Emerging Research Issues and Directions on MaaS, Sustainability and Shared Mobility in Smart Cities with Multi-Modal Transport Systems

Abstract

In recent years, several emerging transport modes have appeared in cities all over the world and have been widely adopted by commuters and travelers. This leads to strong growth and popularity of multi-modal transport and Mobility as a Service (MaaS) in cities. These emerging transport modes have not only received much attention from service providers and practitioners but have also attracted researchers in related communities. These are reflected in the growing number of published papers related to research issues of multi-modal mobility transport in cities. The factors that have been driving the strong growth of the number of published papers related to the emerging multi-modal transport in cities are the deficiencies of effective solution methods to accommodate the needs of users in cities with multi-modal transport modes. Although the existing literature is still deficient in offering seamless end-to-end multi-modal mobility transport services, it provides valuable sources and clues for finding the potential future research subjects/issues/directions. In this study, we attempt to identify potential research directions based on a review of the existing literature on multi-modal mobility transport. By searching the WOS database, we analyze the profile and trends of research directions related to multi-modal mobility. The results of this study pave the way for the assessment of research subjects/issues/directions under the umbrella term of multi-modal mobility transport. This review paper significantly reduces the time required for readers to identify prospective research subjects, issues, or directions without delving into the literature.

1. Introduction

According to the report provided by the United Nations Statistics Division, to combat climate change and its impacts, global carbon dioxide emissions need to be reduced by 45 percent by 2030 from 2010 levels, and reach net-zero emissions by 2050 [1]. Transport accounts for 24% of global CO₂ emissions and three-quarters of these emissions come from road transport [2]. Therefore, reducing the emissions due to road transport is one approach to mitigate and possibly reduce global carbon dioxide emissions. The strong commute and non-commute demand has been driving the growth of the mobility market and increasing the CO₂ emissions. To achieve a more sustainable future, policy makers, transport service providers and practitioners all over the world have been pursuing the Sustainable Development Goals (SDGs) enacted by the United Nations through promoting a variety of emerging shared mobility transport modes and services in cities [3,4]. As a result, various emerging transport modes such as ridesharing, carpooling, carsharing, bike sharing and micromobility have been widely deployed around the world. Furthermore, many interesting and challenging research issues arising from these emerging transport modes lead to studies and review papers related to ridesharing [5,6,7,8,9,10], carpooling [11], carsharing [12,13,14,15,16,17], bike sharing [18,19,20,21,22,23,24,25] and micromobility [26,27,28,29]. These papers, along with their relevance to different emerging transport modes and description, are summarized in

Table 1. Existing studies on shared mobility and micromobility (a related mode is marked by o).

Paper/Year	Ridesharing	Carpooling	Carsharing	Bike Sharing	Micromobility	Description
[5]/2012	o					A review of the literature on the optimization issue of dynamic ridesharing problems based on papers published up to December 2012
[6]/2021	o					A review of ridesharing platforms, user factors and barriers based on 56 articles published between 1990 and 2020
[7]/2020	o					A review of matching, routing, and scheduling methods in peer-to-peer ride sharing based on papers published up to 18 May 2020 (roughly estimated)
[8]/2024	o					Applying a saying to develop self-adaptive algorithms for ridesharing with trust requirements
[9]/2024	o					Comparing hybrid Firefly-Particle Swarm Optimization Algorithm with six hybrid Firefly-Differential Evolution Algorithms for solving ridesharing problems
[10]/2024	o					A success rate based Self-adaptive Differential Evolution algorithm was proposed in [10] for ridesharing systems with a discount guarantee
[11]/2021		o				Analyzing carpooling platforms and proposing a multi-sided platform business model and an architecture for a service provider with multiple customer segments/partners based on articles published up to 18 December 2020
[12]/2022			o			Reviewing studies of the influence of socio-demographic factors on the demand for different carsharing forms based on 91 articles published up to 22 December 2020
[13]/2018			o			Analyzing the different carsharing services and the research questions to classify the research and derive an insight of the mainstreams based on 137 articles published between 2001 and March 2017
[14]/2023			o			A review of mobility-on-demand carsharing systems based on 191 articles published between 2012 and 2021
[15]/2019			o			A systematic review of the vehicle relocation issues in car sharing networks based on 33 articles published between 2010 and 2018
[16]/2023			o			Reviewing the important aspects related to carsharing such as the criteria affecting the development of the carsharing systems, the management of carsharing systems, electro-mobility in carsharing, and service optimization and modeling, based on 60 papers (date range unspecified)
[17]/2024		o	o	o	o	A systematic literature review of the environmental impacts of shared mobility based on 40 papers published up to October 2021
[24]/2021				o		A review of machine learning techniques’ contributions applied to bike-sharing systems to improve cities’ mobility based on 35 articles published between 2015 and 2019
[23]/2021				o		A systemic review of peer-reviewed academic 95 papers between 2016 and 2019 and identifies four key themes in the context of governance of dockless bike-sharing systems
[18]/2020				o		Reviewing the factors affecting station-based bike-sharing demand (date range unspecified)
[25]/2020				o		Reviewing the spatiotemporal characteristics of bike and/or electric bike (e-bike) sharing systems based on a total of 2937 bike sharing systems information assessed in November 2019 to study their distribution, identify the annual growth/expansion of electric bike sharing systems deployment and explore the spatial characteristics in terms of their adoption on a country scale, population coverage and type of system
[19]/2019				o		Reviewing bike sharing research based on 208 articles published between 2010 and 2018 and identifies several bike sharing topic categories such as factors, barrier, safety, health, behavior, impact, sharing economy and system optimization
[20]/2021				o		Focusing on challenges and opportunities of rebalancing in dock-based bike-sharing systems based on articles published between 2010 and September 2020
[21]/2020				o		Reviewing the development, characteristics and impact of bike-sharing systems
[22]/2022				o		A systematic analysis based on the articles related to bike-sharing and 642 papers published between 2010 and 2020 in the Web of Science (WoS) Core Collection was conducted and three categories of topics were identified, (1) operation mode, (2) static rebalancing problem and (3) demand prediction
[26]/2021					o	Providing future research directions related to the role of micromobility in shaping sustainable cities by using a systematic literature review methodology based on 358 papers published between 2000 and 2020
[27]/2021					o	A review regarding use, health and environmental impacts, and policy implications of shared e-scooters based on 70 articles published between 2017 and October 2020
[28]/2023					o	Examining the key determinants, adoption and usage of electric micromobility from a public health perspective based on 67 articles published between 2010 and 2021
[29]/2024					o	A systematic review of the environmental impacts of shared electric micromobility services based on the lifecycle assessment method (20 studies were included in the analysis. They were published between 2019 and 2023)

. However, these studies primarily focus on single transport modes with the exception of [12,13,14,15,16,17], which review the papers related to the impacts of car sharing, carpooling, bike sharing, scooters and moped sharing on the environment published up to October 2021.

The emerging shared mobility transport modes and the existing transport modes coexist to provide multi-modal mobility transport services to users in cities. To provide seamless access to multi-modal mobility transport services for users, the paradigm of Mobility as a Service (MaaS) has emerged. MaaS enables users to plan their itineraries, book the mobility services of different transport modes as needed and pay through a combined platform with multiple heterogeneous mobility modes. Although the vision of MaaS is to facilitate the provision of multi-modal transport services to users by simplifying the procedure to book the required transport services in a bundle based on user friendly GUI without having to book each required transport service one by one, it raises many unaddressed research issues to plan, design and implement MaaS for cities with multi-modal transport modes. These emerging multi-modal transport modes have not only received much attention from service providers and practitioners but have also attracted researchers’ attention in related communities. These are reflected in the number of published review papers related to research issues of MaaS in cities such as [30,31,32,33,34,35,36,37,38,39,40,41,42,43,44], which are centered on Mobility as a Service only. A brief summary of these review papers related to MaaS are listed in

Table 2. Existing studies on MaaS.

Paper	Issue	Year
[30]	To categorize 31 MaaS-focused publications found in Scopus and ScienceDirect databases in June 2018 based on the issues addressed; different transport modes/services in MaaS, MaaS pilots/trials, and the expected effects of MaaS.	2018
[31]	To clarify the meaning of MaaS, the time and place the term MaaS appeared, the main actors in MaaS, how to implement MaaS and the reasons to implement it based on 57 papers published up to January 2019.	2020
[32]	To identify barriers and risks related to adoption of MaaS based on 328 papers published up to 2020.	2021
[33]	To focus on review of whole system simulation methodologies for MaaS based on papers published up to May 2021.	2021
[34]	To provide a literature review on interactions between stakeholders in MaaS based on 50 papers published up to August 2020.	2021
[35]	To provide a literature review of MaaS and public transport based on papers published up to March 2021.	2021
[36]	To highlights three research questions: (1) the topics in the MaaS literature (2) the results in the fields and (3) the research gaps based on 127 publications about MaaS published up to 30 June 2021.	2022
[37]	To review the socio-technical factors for adoption of MaaS and bundling mobility packages based on 29 papers published up to April 2022.	2023
[38]	To review the effects of real-world MaaS applications on socio-territorial inequalities based on 20 papers published before June 2023.	2023
[39]	To draw attention to the MaaS literature that considers gender based on articles published up to 15 March 2023.	2023
[40]	To provide a literature and tool review about personalization in MaaS based on papers published up to March 2023.	2023
[41]	To provide a review on the adoption factors of emerging autonomous vehicles, drones, micromobility and Mobility as a Service, based on papers published up to June 2021.	2023
[42]	To review 31 papers on MaaS models articles published up to February 2022 and draw the conclusion that most studies were demand-centered usage models without considering the integration aspect of the system’s supply side’s and related operational and profitability factors.	2024
[43]	To provide a review of cyber security risks in MaaS ecosystems and suggests a research agenda based on articles published up to June 2024.	2024
[44]	To propose a nested ecosystem framework involving actors, infrastructure, value, and customers to guide both transport sector academics and practitioners and highlight the challenges in the deployment of MaaS based on 125 articles published up to July 2024.	2024
[45]	To provide a systematic literature review of sustainable transport in urban traveling based on 84 articles published between 2014 and 2024 and propose an integrated framework to develop cooperation-oriented multi-modal shared mobility.	2024

.

The factors that have been driving the strong growth of the number of published papers related to the emerging multi-modal transport in cities are the deficiencies of effective solution methods to accommodate the needs of users with multi-modal transport modes. Although the existing literature is still deficient in offering seamless end-to-end multi-modal mobility transport services, it provides valuable sources and clues for finding the life cycle of research subjects, issues and future directions. The goal of this study is to identify potential research directions based on analyzing the profile and trends of the recent multi-modal mobility transport literature obtained by searching the WOS database. The results of this study would pave the way for the assessment of research subjects, issues and directions related to MaaS, sustainability and shared mobility in the context of multi-modal mobility transport. This review paper significantly reduces the time required for readers to identify prospective research subjects, issues, or directions without delving into the literature.

This paper is different from the existing review papers on ridesharing [5,6,7], carpooling [11], carsharing [12,13,14,15,16,17] and bike sharing [18,19,20,21,22,23,24,25] and micromobility [26,27,28,29] in Table 1. The review papers mentioned above focus on the research issues for systems with a single transport mode, whereas this paper focuses on the research issues for systems with multiple transport modes. Since research on multi-modal mobility transport is still in its infancy, this review will limit its scope to papers published between 2017 and 2024.

Although this paper focuses on research issues related to multi-modal mobility and covers MaaS, it is different from the ones in Table 2. Note that most of the studies included in the review papers listed in Table 2 were published before June 2023 with the exception of [43,44,45]. It is necessary to review more recent articles to understand the trends. Moreover, although the paper [44] includes 125 papers published in the timeframe 2014 to July 2024 in the review and provides some discussions and future research directions, the research questions/issues, results/contributions and limitation of each of the 125 papers are not explicitly summarized. Although the paper [45] includes 84 papers published in the timeframe 2014 to 2024 in its review, identifies four themes of shared mobility for sustainable transport in the context of urban traveling and provides an integrated framework to facilitate the development of cooperation-oriented multi-modal shared mobility, concrete future research directions, questions and issues related to the integrated framework are not explicitly summarized or described. The current paper is different from the review papers in Table 2 in terms of timeframe as it focuses on papers published in the time frame January 2017 to December 2024. More importantly, the research questions/issues, results/contributions, limitations of each of the papers and future research directions are explicitly described and discussed in this review.

The rest of this paper is organized as follows. In Section 2, the method used to perform this survey is summarized. This includes the details of the keywords and query strings used to search the papers relevant to research subjects in multi-modal mobility. In Section 3, the papers related to MaaS in the context of multi-modal mobility will be reviewed. In Section 4, the papers related to sustainable multi-modal mobility will be reviewed. In Section 5, the papers related to shared mobility or ridesharing in the context of multi-modal mobility will be reviewed. Section 6 will focus on the review of the papers related to micromobility in the context of multi-modal mobility. Section 7 will focus on the review of the papers related to the integration issue in the context of multi-modal mobility. Discussions based on the results of this review will be conducted in Section 8. We will conclude this study in Section 9.

2. The Method and Materials

The method used in this study can be broken down into several steps as shown in Figure 1. The source for searching the candidate papers for review in this study is the Web of Science Core Collection. The steps include the initial search stage to acquire the papers that might be relevant to the issues of multi-modal transport modes. As we focused on issues related to multi-modal mobility, the keywords used to search the candidate papers from the Web of Science Core Collection for review in the initial search were “mobility” and “multi-modal”. The results of the initial search were exported in the Tab Delimited File format so that VOSviewer can be used to analyze the results. The results obtained by VOSviewer were used to identify terms or keywords to refine the search. We categorized the search by combining different terms or keywords to perform refined search. We refined the search based on different combinations of terms or keywords find the works in different categories. Finally, we analyzed the contents of the works in each category to identify the research questions addressed, results or findings and potential research directions.

2.1. The Initial Search with Keywords ‘Mobility’ and ‘Multi-Modal’

In this paper, we review the literature relevant to different types of transport modes in smart cities. The sources of the literature are obtained based on the Web of Science database. As this paper focuses on the studies relevant to multi-modal mobility in smart cities, the keywords used in the initial search are “mobility” and “multi-modal”. The query string we used is AB = ((“mobility” AND “multi-modal”)). The search returned 225 results from the Web of Science Core Collection. The distribution of published papers in the search results from 1996 to 2024 is shown in Figure 2. The number of published papers during the years 1996 to 2005 is 6. The number of published papers during the years 2006 to 2015 is 34. The number of published papers during the years 2016 to 2024 is 185. The number of published papers has grown exponentially over the years.

Figure 3 shows the network visualization in VOSviewer for the initial search. Note that the keywords in the items that appear in the network visualization such as multi-modal, sustainability, transportation, ride, sharing and bike provide the clue for refining the search to find several categories of works related to multi-modal mobility.

The items detected in the network visualization were used to refine the search. In the remainder of this paper, we classify the papers into several categories by refining the search through different combinations of keywords.

2.2. The Refined Search

As we are interested in the recent research development, the date range for the search is from January 2017 to December 2024. The search resulted in over 189 papers. To identify the emerging research directions in the past 8 years, we refined the search by changing the query string in several ways to obtain published papers related to different issues or sub-categories.

The items detected in the network visualization by VOSviewer were used as keywords for refining the search. In addition, as integration and pricing are two important issues in MaaS, “MaaS” is also combined with “integration” or “pricing” in the refined search. To categorize the works, we refined the search by the following combinations of keywords:

(i)

(“MaaS” AND “multi-modal”);

(ii)

(“sustainable” AND “mobility” AND “multi-modal”) OR (“sustainability” AND “mobility” AND “multi-modal”);

(iii)

(“shared mobility” AND “multi-modal”) OR (“ridesharing” AND “multi-modal”);

(iv)

(“bike” AND “transport” AND “multi-modal”) OR (“micromobility” AND “multi-modal”) OR (“scooter” AND “transport” AND “multi-modal”);

(v)

(“MaaS” AND “integration”) OR (“MaaS” AND “pricing”).

As the goal of this paper is to find the potential research directions, we narrowed the search scope by setting the publication date range from 1 January 2017 to 31 December 2024. By searching the Web of Science Core Collection with the query string AB = (“MaaS” AND “multi-modal”) for the papers published between 2017 and 2024, we found 18 papers. By searching the Web of Science Core Collection with the query string AB = (“sustainable” AND “mobility” AND “multi-modal”) OR AB = (“sustainability” AND “mobility” AND “multi-modal”), we found 31 papers. By searching the Web of Science Core Collection with the query string AB = (“shared mobility” AND “multi-modal”) OR AB = (“ridesharing” AND “multi-modal”), we found 13 papers. By searching the Web of Science Core Collection with the query string AB = ((“bike” AND “transport” AND “multi-modal”) OR (“micromobility” AND “multi-modal”) OR (“scooter” AND “transport” AND “multi-modal”)), we found 17 papers. By searching the Web of Science Core Collection with the query string AB = (“MaaS” AND “integration”) OR AB = (“MaaS” AND “pricing”), we found 96 papers.

As the papers obtained from the research may not be of interest to multi-modal, content analysis was performed to identify the topics covered in each paper. Through content analysis, we identified several topics, including adoption factors, modeling, simulation and planning, impact, implementation, integration and performance.

Table 3. Papers included in this review and the associated topics.

Topic	Papers
Adoption factors	[46] (2021), [47] (2022), [48] (2023), [49] (2020), [50] (2022), [51] (2021), [52] (2020), [53] (2017), [54] (2019), [55] (2020), [56] (2020), [57] (2019), [58] (2023), [59] (2024), [60] (2017), [61] (2018), [62] (2024), [63] (2024), [64] (2024)
Modeling, simulation and planning	[65] (2023), [66] (2022), [67] (2023), [68] (2021), [69] (2019), [70] (2017), [71] (2019), [72] (2018), [73] (2024), [74] (2019) [75] (2023), [76] (2018), [77] (2018), [78] (2020), [79] (2024), [80] (2020), [81] 2023, [82] (2019), [83] (2021)
Impact	[84] (2022), [85] (2017), [86] (2024), [87] (2023), [88] (2023), [89] (2023), [90] (2022), [91] (2021), [92] (2018), [93] (2023), [94] (2021), [95] (2023), [96] (2024)
Implementation	[97] (2019), [98] (2020), [99] (2023), [100] (2018), [101] (2022), [102] (2020), [103] (2022), [104] (2021), [105] (2022), [106] (2020)
Performance	[107] (2022), [108] (2020)
Integration	[109] (2021), [110] (2020), [111] (2020), [112] (2020), [113] (2018), [114] (2024), [115] (2020), [116] (2020), [117] (2024), [118] (2021), [119] (2020), [120] (2024), [121] (2024), [122] (2024), [123] (2024)

summarizes the topic of each paper in this review. Note that Table 3 only summarizes the topic for each paper. The details regarding the research questions, results/contributions and gaps of the reviewed papers will be presented in the tables of Section 3, Section 4, Section 5, Section 6 and Section 7. In-depth discussions related to the emerging research directions are included in Section 8.

3. Review of Studies Related to MaaS in the Context of Multi-Modal Mobility

Mobility as a Service (MaaS) enables users to seamlessly access various modes of transport and related services such as trip planning booking and payment operations via a single, comprehensive, and on-demand mobility service instead of multiple ticketing and payment operations. In this section, we will present the details to find the papers related to MaaS in the context of multi-modal mobility, review them and identify potential future research directions.

3.1. Summary of Studies

The term “MaaS” in conjunction with “multi-modal” provides a clue for narrowing down the number of papers relevant to both “MaaS” and “multi-modal”. The query string we used to search the first category of papers is AB = (“MaaS” AND “multi-modal”), which is aimed to find the papers with keywords “MaaS” and “multi-modal” in the abstract. The search returned 18 results from the Web of Science Core Collection. Figure 4 shows the distribution of these papers published over the years. After content analysis, 17 papers are retained. These papers were published during the years between 2017 and 2025. Five papers were published during the years between 2017 and 2020 and 12 papers were published during the years between 2021 and 2024. The number of published papers increases during the years from 2021 to 2023. Based on the above information, papers with the keywords “MaaS” and “multi-modal” in the abstract might provide a clue for identifying potential emerging research directions.

The research questions addressed in the papers found by the refined search of this section are summarized in

Table 4. Summary of research questions addressed in the papers published between 2017 and 2024 in the Web of Science Core Collection with keywords “MaaS” and “multi-modal” in the abstract.

Paper	Research Questions
[46]	What are the challenges and opportunities of sustainable mobility in post-COVID-19 period?
[47]	Is MaaS able to address the challenges of mothers in multi-modal mobility?
[65]	How to develop a general modelling framework for MaaS based on analysis of the relevant literature considering main actors and factors?
[97]	What are the requirements of MaaS? How to develop a sustainable business model for transport operators at varying levels of MaaS considering risk and data sharing?
[48]	Who and why users took part in the trial of a MaaS service? Whether the trial experience satisfied the users’ motives?
[84]	How MaaS could impact GHG emissions in an urban setting?
[66]	How to simulate multi-modal operations of Mobility as a Service in low-density cities?
[49]	1. To what extent can the MaaS promises to people and government be delivered? Could any unanticipated societal issues arise from adoption of MaaS?2. What are the challenges for urban governance from potential MaaS adoption, and the recommended responses to these challenges?
[50]	Are there principles for sustainable Mobility as a Service? Can we identify principles for sustainable MaaS in areas with high car dependency?
[51]	How does MaaS influence commuting mode choice of commuters?
[98]	How to implement multi-modal shared mobility through policy implementation?
[99]	What is the planning process for implementing MaaS?
[100]	How to develop an architecture for MaaS to tackle the potential threats?
[85]	Whether the way bus services might be offered will be changed under MaaS?
[67]	How to design a mobility hub to achieve sustainability, performance and efficiency by reducing vehicle miles in the city and combining many multi-modal transport options for transferring passengers?
[86]	How about the competitiveness of MaaS multi -modal travel options with respect to private car usage?
[68]	How to determine the route choices of different transport modes that can jointly form a complete path for transporting the user from origin to destination?

.

3.2. Review of Studies

The transportation problems in cities with multi-transport modes can be modeled as “System of Systems” design problems [124]. A MaaS system is an example of a “System of Systems” in which different modes of transport systems work together with the goal to provide seamless end-to-end services to transport users. Due to the existence of multiple transport modes, there are many types of stakeholders involved in the management and operations of the MaaS system. These stakeholders include the transportation planners, service providers of different transport modes, MaaS providers and users or travelers of transport services. The transportation planners are responsible for suggesting the proper regulations to be made and adopted to enable efficient, safe and sustainable transport services and ensure that companies, organizations and individuals abide by them to promote transport market through healthy competition. In cities with multiple heterogeneous transport modes, there are typically multiple single transport mode service providers which provide single transport mode services to users or travelers within their service area. As the itineraries of users or travelers may not be covered in the service area of a single transport mode service provider, multiple heterogeneous transport modes are needed to fulfill the requirements of users or travelers. Without a MaaS system, users or travelers requiring multiple heterogeneous transport modes have to book and pay the tickets through multiple single transport mode service providers one by one. The role of MaaS providers is to improve the efficiency to plan, book and pay the tickets for the multiple heterogeneous transport modes needed according to the itineraries of users or travelers. The role of users or travelers is to plan, book, pay and use the tickets for the required multiple heterogeneous transport modes according to their itineraries by following the rules from the service providers of different transport modes and the transportation planners.

Depending on the type of stakeholders mentioned above, the concerns or questions about taking part in the operations of MaaS vary. The concerns or questions of each type of stakeholders can be linked to related research issues or topics for further studies. For example, from the perspective of MaaS providers, capturing the heterogeneous travel needs of users’ and the interactions between users’ choices in planning and booking tickets for the multiple heterogeneous transport modes is an important issue. Another example, from the perspective of mothers with babies, as strollers are often accompanied with mothers, this may pose a challenge for mothers to access stations of multiple transport modes in MaaS and move from one transport mode to another. From the perspective of the government, transportation planners may want to understand how MaaS could impact GHG emissions in an urban setting and improve sustainability. For single mode transport service providers such as traditional bus operators, they may want to know whether MaaS will change how bus services are offered.

This subsection focuses on the review of existing studies in Table 4 related to MaaS in the context of multi-modal mobility. The research questions along with the corresponding study results are summarized as follows.

(1)

Question: What are the challenges and opportunities of sustainable mobility in post-COVID-19 period?

Results: The study [46] is about the challenges and opportunities of sustainable mobility in post-COVID-19 period. The study showed that one in five car-less households intended to purchase a car during COVID-19 period due to the need of physical distancing, convenience, safety, low traffic levels and low parking demand. The authors projected that a regional scale MaaS program may be created due to the widespread available real-time crowding information in mass transit systems and apps provided by transport service providers.

(2)

Question: Is MaaS able to address the challenges of mothers in multi-modal mobility?

Results: In [47], the authors attempted to answer the question “Does MaaS address the challenges of multi-modal mothers?”. The authors identified the needs of mothers from their perspectives to answer this question. By analyzing MaaS and comparing MaaS with the existing system from the perspectives of transportation, technology, policy, subject and economics, the study found that the needs of participants include conveniently accessible stations, tram with strollers and infrastructure. The authors suggested that offering different micromobility vehicles such as shared cargo bikes or electric bikes and the functions to plan and make an advance reservation can be beneficial for mothers.

(3)

Question: How to develop a general modelling framework for MaaS based on analysis of the relevant literature considering main actors and factors?

Results: Due to the lack of a model to fully represent the diverse travel needs of users’ and aspects of the interaction between choices, a generic framework need to be proposed to model MaaS. In the paper [65], a generic framework considering the factors of (i) socio-economic characteristics, (ii) attitudes and habits of the travelers, and (iii) MaaS-related factors has been proposed to model MaaS to capture users’ diversified needs and the service providers’ conflicting objectives.

(4)

Questions: What are the requirements of MaaS? How to develop a sustainable business model for transport operators at varying levels of MaaS considering risk and data sharing?

Results: There lacked a study on the requirements of a sustainable business model for transport operators at different levels of MaaS considering risk and data sharing. The authors of [97] proposed a MaaS Lite model to focus on the users and the end-to-end trips. The MaaS Lite model aims to deliver a user-centric multi-modal mobility services and overcome many barriers to implement MaaS such as institutional inertia, poor scalability and lack of trust.

(5)

Questions: Who and why users took part in the trial of a MaaS service? Whether the trial experience satisfied the users’ motives?

Results: To understand who and why users took part in the trial of a MaaS service in Sydney and whether the trial experience satisfied their motives, analysis based on questionnaires and interviews was performed in [48]. The study of [48] indicates the conjecture that MaaS does not appeal to private car users is not correct, which implies that multi-modal travelers are potential trend setters of MaaS. The study found that the main motive to take part in the trial is a wish to contribute to innovative MaaS services and curiosity about the new multi-modal transport.

(6)

Question: How MaaS could impact GHG emissions in an urban setting?

Results: The question about the impact of MaaS on GHG emissions is of interest to the government transportation planners as it is responsible for introducing and promoting sustainable transport. In [84], the authors studied how MaaS could impact GHG emissions in an urban setting. Three scenarios of how MaaS could impact GHG emissions in an urban setting were studied based on activity-based model for Amsterdam, the Netherlands. The study indicated that the implementation of MaaS might lead to emission reductions. The attractiveness of MaaS will largely determine the emission reduction potential.

(7)

Question: How to simulate multi-modal operations of Mobility as a Service in low-density cities?

Results: Due to the shortcomings of traditional transportation planning without considering the aspects of adoption, partnerships, capacity, operations, integration, and impact on multi-modal planning of low-density cities, simulation of multi-modal operations is required. A simulated MaaS network was used in the study [66] to evaluate a variety of metrics and inter-modal interactions based on experiments in International Drive in Orlando, Florida. The experiments were performed based on scenarios of several transport modes, including walking, micromobility, private vehicles, ridesourcing and transit, with varied modal shares under a variety of congestion conditions. The analysis of results shows the strengths and weaknesses of transport modes. Transit enjoys the lowest impact per person. Ridesourcing has adverse effects across different performance measures. The study outlines strategies to reduce travel time, waiting time, and overall transport network delay based on optimization of ridesourcing–transit integration.

(8)

Questions: To what extent can the MaaS promises to people and government be delivered? Could any unanticipated societal issues arise from adoption of MaaS? What are the challenges for urban governance from potential MaaS adoption, and the recommended responses to these challenges?

Results: Despite the fact that many studies show that MaaS has the potential benefits for users by providing efficient access to plan and book services of multiple transport modes to meet their transport requirements through one application, some researchers are skeptical about what MaaS promised to people and government. The paper [49] attempted to answer the question about the extent to which the MaaS promised to people could be delivered and studied the challenges for urban governance. The authors assessed the grand promises that MaaS has made in terms of efficiency, choice and freedom. The authors concluded that MaaS has considerable potential for deception due to the unanticipated societal issues that could arise from adoption of MaaS. In addition, without proper government intervention, the MaaS efficiency and equity promises are not possible.

(9)

Questions: Are there principles for sustainable MaaS? Can we identify principles for sustainable MaaS in areas with high car dependency?

Results: Although MaaS has been proposed as a potential solution to deal with the economic, environmental and social issues arising from car dependency, there still lacks a study on the principles for sustainable MaaS. The paper [50] aims to identify principles for sustainable MaaS in the context of unsustainable car dependency. Several principles for sustainable MaaS were set out in [50] by reviewing recent literature to help determine the principles. The principles take into consideration several factors, including (1) the need to reduce the reliance on using a personal vehicle for transportation, (2) the need for access to workplaces and amenities, (3) the need to minimize the environmental impact and (4) the need to be long-term viable for all stakeholders. MaaS operators are recommended to design services by following these principles.

(10)

Question: How MaaS influences commuting mode choice of commuters?

Results: MaaS provides end-to-end transport services to users by combining multi-transport modes through applications on mobile phones. Commuters must determine commuting mode choice when MaaS is used. How MaaS influences commuting mode choice of commuters is an important issue. To investigate this issue, several tasks have been performed in the study [51], including (1) examine the potential role of MaaS as a transport demand management tool, (2) conduct a stated choice experiment among employees in the Netherlands, (3) show that MaaS could be a promising transport demand management tool when combining carrots (low MaaS package prices) and sticks (parking price increases), and (4) examine the behavioral change potential per user category.

(11)

Question: How to implement multi-modal shared mobility through policy implementation?

Results: The widespread adoption of smart devices with geographic positioning system, geographic information systems, software apps, capability of electronic transactions and increasing availability of urban mobility options through multi-modal transport systems and shared mobility enable the emerging MaaS paradigm. Although the number of related papers has increased significantly, there is still a lack of reviews of the policy for the implementation of MaaS. In [98], policy implementation of multi-modal shared mobility considering both the supply and demand sides has been reviewed. The results showed that providing car sharing, bike sharing, and e-scooter sharing for first/last mile access to transit systems from the supply side can improve transport accessibility and reduce social inequality from the demand side. Implementation of multi-modal shared mobility calls for a collaborative partnership of service providers and government to jointly set a merit-based business model to improve infrastructure provision, and create smart applications.

(12)

Question: What is the planning process for implementing MaaS?

Results: Despite many MaaS pilot projects have been executed in different cities, most related studies focus on the barriers and travel behaviors of MaaS users. There lacks a study on the planning process for the implementation of MaaS. The paper [99] reports the planning process for integrating the services of several transport operators, including public transport operators, a micromobility service provider and a taxi operator, into the MaaS demo site of Athens’ in Greece. The planning of three travel cases was covered in this study: (1) multi-modal trips for work; (2) MaaS trips for tourists; and (3) interurban trips for work and shopping/leisure.

(13)

Question: How to develop an architecture for MaaS to tackle the potential threats?

Results: MaaS relies on federations of transport service providers. Providers trade their resources to attempt to create multi-modal mobility solutions through coordination. Although MaaS may help generate multi-modal solutions for mobility flexibly, it comes with insider threat and security/privacy concerns. In addition, the emerging networks of ubiquitous, pervasive devices providing real-time data on objects and people called Cloud-of-Things for mobility provide potential attack surface for insiders. A study on the potential threat in MaaS is required. The study of [100] focuses on security and privacy issues from the perspective of an insider threat. The study classifies the potential threats into two types: the potential threats to individual operators and the potential threats to federated MaaS providers. Appropriate measures to mitigate the problems were proposed. An overlay networking architecture was proposed to constrain the data obtainable by insiders while optimizing the users’ requests.

(14)

Question: Whether the way bus services might be offered will be changed under MaaS?

Results: The potential benefits of MaaS are expected to have a profound impact on traditional transport services such as bus services. An important research question for traditional bus service providers is whether the way bus services might be offered will be changed under MaaS. This research question motivated the study reported in [85]. The study indicates that a new hybrid multi-modal contract setting driven by the mode-neutral customer experiences might change the way bus services will be offered under MaaS.

(15)

Question: How to design a mobility hub to achieve sustainability, performance and efficiency by reducing vehicle miles in the city and combining many multi-modal transport options for transferring passengers?

Results: A well-designed mobility hub is vital for reducing vehicle miles in the city and provides accessible space as well as many multi-modal transport options for transferring passengers. Design of a mobility hub is an important issue to achieve sustainability, performance and efficiency. In [67], the authors proposed the methodology and analyzed the parameters for designing process of mobility hubs. Design parameters for a new mobility hub center are determined via expert opinions and literature review. The authors proposed an integrated methodology that consists of Fermatean fuzzy sets, Quality Function Deployment and visualization of mobility hub center.

(16)

Question: How about the competitiveness of MaaS travel options with respect to private car usage?

Results: One approach to assessing the potential for MaaS to reduce the use of private cars in cities is to study the competitiveness of MaaS with respect to private car usage? The study [86] analyzes the competitiveness of MaaS versus private car usage for the commuting problem in a one-origin–one-destination network based on consideration of fixed cost for travel and the inconvenience cost of multi-modal trips. The analysis indicates that mode choice is affected by fixed and inconvenience cost and depends on demand. The entry fee to adopt different modes is set by fixed costs. The transportation resources utilization is influenced by inconvenience costs. The results suggest employing pricing and capacity strategies to maintain a balanced mode share.

(17)

Question: How to determine the route choices of different transport modes that can jointly form a complete path for transporting the user from origin to destination?

Results: As a user provides his/her transportation request with given origin and destination to the MaaS decision support system, the system must determine the route choices of different transport modes that can jointly form a complete path consisting of multiple route choices for transporting the user from origin to destination. However, there still lacked a study on this problem. This problem was addressed in [68] as the generalized path overlapping problem described by a multi-modal logit kernel model to take into account the correlations of unobserved utilities of combined routes. The stochastic equilibrium on multi-modal transport networks was formulated as a fixed-point problem base on this model. It was found that the proposed solution method significantly affects both the route choice predictions and traffic flows at equilibrium. The numerical studies of this study were performed based on a novel network.

3.3. Potential Future Research Directions

The above review of the recent papers related to MaaS in the context of multi-modal mobility shows that there are various facets in terms of research issues or questions. These research issues or questions include (i) enabling factors/barriers of modal shift to sustainable transport, (ii) analysis of the potential impact caused by MaaS, (iii) strategies to facilitate the implementation of multi-modal shared mobility, (iv) advanced frameworks/models to simulate/optimize the performance of MaaS and (v) effective approaches to dealing with threats while protecting privacy and allowing data sharing in MaaS. Several potential future research directions based on the review of the papers in the previous subsection are summarized in this subsection.

(i)

Enabling factors/barriers of modal shift to sustainable transport

The study [46] identified the challenges and opportunities of sustainable mobility in post-COVID-19 period and projected that MaaS might be created due to the widespread available real-time crowding information in mass transit systems and apps. To enable a shift away from cars, the study [46] suggests several future research directions, including multi-modal coordination, flexible choice of a variety of transport modes, congestion pricing and market-rate parking policies. To benefit from MaaS, the study of [47] mentioned that the network of different transport modes (e.g., the public transportation network, the cycling network and the pedestrian network), must be complete in itself. The study [50] attempted to identify principles for sustainable MaaS in the context of unsustainable car dependency. The study also pointed out several application scenarios not explored in the literature: (1) How to deal with the situation of not having access to a car? (2) Is it possible for a multi-modal journey with luggage? (3) Does the service support families with children? (4) Can it provide access for those with mobility impairments? (5) Does it provide service for people without a smartphone?

The work [48] attempted to find the answer to the questions: (1) who and why users took part in the trial of a MaaS service in Sydney and (2) whether the trial experience satisfied the users’ motives. As the analysis of the study [48] was based on the trial service prototype of a MaaS, further studies considering changes to design of services, target group, or relevant context that may significantly influence the users’ experiences are required to generalize the findings to all MaaS operations.

(ii)

Analysis of the potential impact caused by MaaS

The study [84] probed into the question: how MaaS could impact GHG emissions in an urban setting? As the study [84] was based on Amsterdam, the Netherlands, a future research direction is to investigate to what extent the conclusions made still hold. The emerging MaaS relies on integration of different transport modes as well as cooperation of transport service providers to improve efficiency, performance and sustainability. Therefore, MaaS would have impact on existing transport modes. The work [85] investigated whether the way bus services might be offered will be changed under MaaS. The study [85] suggested several future research directions: (1) investigating the extent to which individuals would be willing to not own a car, (2) under what conditions individuals owning cars would be willing to offer their cars for use by others, and (3) how much travelers are willing to pay for improving the journey experience. The study [86] focused on the competitiveness analysis of MaaS travel options with respect to private car usage. in a one-origin–one-destination network based on consideration of fixed cost for travel and the inconvenience cost due to multi-modal trips. Incorporation of the factors of transit frequency, capacity of vehicles and hidden costs associated with the comfort level during trips in the analysis requires a further study. Development of a management scheme to optimize the system is a challenging issue.

(iii)

Strategies to facilitate the implementation of multi-modal shared mobility

Although MaaS may enable cities to achieve more sustainable future, an important issue is how to implement multi-modal shared mobility through policy implementation. The study [98] called for the government policy makers to innovatively promote shared MaaS. Establishment and operation of shared MaaS requires the development of Merit-based business models, organizational promotions, provision of infrastructure, and smart operation systems. The paper [99] reported the planning process for integrating the services of several transport operators, including public transport operators, a micromobility service provider and a taxi operator, into the MaaS demo site of Athens’ in Greece. One future research direction is to use surveys to focus on desired app features. Developing a mobility platform to facilitate combination of transport modes and the business models to promote such combination and guiding the interested stakeholders to create sustainable MaaS schemes are also potential future research directions. The work [97] analyzed the requirements of MaaS and the development of a sustainable business model describing connectivity between transport operators for MaaS considering risk reallocation and data sharing. A future research direction suggested in [97] is to standardize interfaces with operators to extend the proposed MaaS Lite model to achieve operational integration.

(iv)

Advanced frameworks/models to simulate/optimize the performance of MaaS

The work [67] focuses on design of a mobility hub through combining many transport options for transferring of passengers and reducing vehicle miles in the city to achieve sustainability, high performance and efficiency. There are two future research directions following the study [67]: (1) extending the proposed method by integrating interval-valued Fermatean fuzzy sets to improve robustness and reliability of the proposed framework and (2) studying the influence of the number of decision making on the outcomes.

In the work [65], Cisterna et al. studied how to develop a general modelling framework for MaaS based on analysis and consideration of main actors and factors. The model proposed in [65] considers the first generation of MaaS. Further study on integrating non-transport features into the model for the second generation of MaaS is required. Development of microsimulation models based on agents is an important future research direction. The development of multi-actor models for MaaS service providers, MaaS Broker and the government to adapt to the specific area is another interesting research direction.

To evaluate a variety of performance measures and inter-modal interactions, the study [66] focused on the issue to simulate multi-modal operations of MaaS in low-density cities. The study outlined strategies to reduce travel time, waiting time, and overall network delay based on optimization of ridesourcing–transit integration. The study [66] pointed out that further study on optimization of ridesourcing–transit integration is required.

The problem to determine the route choices of different transport modes for transporting the user from origin to destination in MaaS was addressed in the study [68]. Assessment of the prediction performances and comparison with other models based on the stated preference surveys and smart phone data is an important future research direction following the study [68]. Studies on matching the trajectory data to the multi-modal network and the generation of choice set for combined routes are also important future research directions.

(v)

Effective approaches to dealing with threats while protecting privacy and allowing data sharing in MaaS

The study of [100] focused on security and privacy issues and proposed an overlay networking architecture to constrain the data obtainable at the data sharing and processing level while optimizing the users’ requests. The authors suggested two future research directions: (1) to implement a mechanism to formally verify a service based on workflow and acceptable behaviors for each service category above the data sharing and processing level; (2) a further study on an architecture to mitigate distributed denial-of-service attacks.

4. Review of Studies Related to Sustainability in the Context of Multi-Modal Mobility

Transport accounts for one-fourth of energy-related emissions [3,4] and is a major driver behind the growing global energy demand [125]. Therefore, countries worldwide have been promoting sustainable transport to achieve Sustainable Development Goals (SDGs). In this section, we will present the details to find the papers relevant to sustainable mobility in the context of multi-modal mobility, review them and identify potential future research directions.

4.1. Summary of Studies

The keyword “sustainability” detected in Figure 3 in conjunction with “mobility” and “multi-modal” provides a clue for narrowing down the number of papers that are relevant to “MaaS” and “multi-modal”. As the keyword “sustainable” is also commonly used as an alternative in the literature, we used two query strings AB = ((“sustainable” AND “mobility” AND “multi-modal”)) and AB = ((“sustainability” AND “mobility” AND “multi-modal”)) to search the papers related to sustainable mobility with multi-modal transport. The sources of the literature were obtained based on the Web of Science database. The refined search resulted in 31 papers. After performing the search and content analysis, 19 papers were retained. Figure 5 shows the distribution of these papers published over the years. Although the number of papers published between 2021 and 2024 (7 papers) is smaller than the number of papers published between 2017 and 2020 (12 papers), there are new emerging research issues. This indicates that the sustainability is still an important issue in the context of multi-modal mobility.

The research questions addressed in each paper found by the refined search of this section are summarized in

Table 5. Summary of research questions addressed in the papers published between 2017 and 2024 in the Web of Science Core Collection with keywords “sustainable/sustainability”, “mobility” and “multi-modal” in the abstract.

Paper	Research Questions
[52]	What are the determinants of mobility consumers’ purchase intentions for mobility products with multi-modal transport modes?
[107]	How to estimate long-term travel delay in multi-modal systems with different transport modes such as subway, taxi, bus, and personal cars?
[101]	How to develop a decision framework for supporting integrated sustainable urban mobility policy design?
[69]	How to develop an appropriate model to support multi-modal travel planning and implementation of the transportation infrastructures?
[70]	How to develop a framework for modeling various facets of Automated Mobility on Demand (AMoD) jointly taking into account the decision making of drivers and the centralized coordinated fleet management?
[53]	Is there a connection between life course and sustainable mobility for the millennial generation?
[87]	What is the impact of the COVID-19 pandemic on public transportation system?
[54]	Is it possible to encourage people to make sustainable mobility choices and reduce car dependency/emissions and energy consumption through the development of a smartphone app?
[88]	What is the effect of sustainable mobility strategies on reducing the use of private vehicles?
[71]	How to develop a framework for modeling traveler behaviors and incentivize travelers to select more sustainable travel options in on-demand mobility services?
[55]	Is there a connection between transport mode choice and travel satisfaction? Is it possible to identify the connection between transport mode choice and travel satisfaction?
[56]	What are the factors that drive the adoption of ride-hailing and the associated travel characteristics? Are there mode substitution effects for ride-hailing?
[72]	How to construct a simulation model to assess the potential impact of automated mobility-on-demand on urban mobility such as its disruptive effect on urban transportation for travelers currently using mass transit or private vehicles?
[57]	Do the characteristics and motives differ between business-to-consumer carsharing adopters and peer-to-peer carsharing adopters?
[73]	How to develop a model that incorporates various sustainability factors to provide decision support for sustainable household mobility choices in a multi-modal transport network?
[58]	How to design a scheme to reward travelers who have adopted sustainable transport modes?
[108]	How to determine and analyze the patterns of the modal accessibility gap in cities?
[89]	How to assess the impact of home-to-work travel on sustainability?
[102]	How to reconstruct a damaged urban system following earthquake to implement a more sustainable efficient urban transport system?

.

4.2. Review of Studies

The pressure to achieve sustainability development goals has created several interesting research issues or questions that pose challenges in the development of approaches to identifying the factors that decisively affects the adoption of various transport modes and sustainable mobility behaviors by mobility users, creating effective strategies to support the design of urban mobility policy, planning/implementing transportation infrastructure for sustainable transport, encouraging/incentivizing travelers to select more sustainable travel options and assessing the impact of emerging transport modes and strategies on sustainability solutions. These research questions along with the studies that respond to them are summarized as follows.

(1)

Question: What are the determinants of mobility consumers’ purchase intentions for mobility products with multi-modal transport modes?

Results: To find the determinants for travelers to adopt sustainable mobility behaviors, the paper [52] focused on study of mobility motives of travelers. The authors conducted a survey experiment to collect potential mobility consumers’ mobility motives and their purchase intentions for diverse multi-modal mobility products. The results indicate that mobility motives considerably contribute to explaining all purchase intentions. Status, financial, independence, safety and hedonic motives contribute substantially to predicting mobility purchase intentions.

(2)

Question: How to estimate long-term travel delay in multi-modal systems with different transport modes such as subway, taxi, bus, and personal cars?

Results: One challenge in multi-modal transport is to estimate the long-term multi-modal travel delay. This arises from the multiple trips across different transport modes and the uncertainty for transferring from one mode to another in the multi-modal travel. The study [107] focused on measurement of long-term multi-modal travel delay based on a 5-year dataset of 8 million residents from 2013 to 2017 in the Chinese city Shenzhen. It was found that (1) the increase in delay of aboveground system is higher than that of the underground system; (2) the upgrade of the underground system infrastructure decreases the increase in the aboveground system travel delay; (3) the underground system travel delays decrease for region with higher population during the peak hours.

(3)

Question: How to develop a decision framework for supporting integrated sustainable urban policy design?

Results: The requirements of designing environmentally sustainable, economically viable, and socially acceptable urban mobility systems pose challenges in the decision-making processes of cities. The issue is to develop an innovative and collaborative decision framework to support integrated sustainable urban policy design for sustainable urban mobility. In [101], the authors proposed a multi-level decision framework for the design of urban mobility policies based on a multi-criteria resource allocation decisions at different levels. To support decision makers, an interactive scenario-simulation tool was developed. The proposed model was applied to the metropolitan region of Rhine-Ruhr for verification.

(4)

Question: How to develop an appropriate model to support multi-modal travel planning and implementation of the transportation infrastructures?

Results: Combining public transport with other transport modes such as shared and on-demand systems results in increasing options in multi-modal travel, leads to increase in complexity and poses a challenge in planning and implementing the transportation infrastructures. Developing an appropriate model to support planning and implementing the transportation infrastructures is important. The paper [69] focused on the implementation of modality switch points (MSPs) within the road network, where MSPs indicate the locations along the route at which travelers can switch from one transport mode to the next. In [69], a model was presented to allow agents to switch between different transport modes at any time to meet the requirements of versatile, individual, and spontaneous multi-modal travel.

(5)

Question: How to develop a framework for modeling various facets of (Automated Mobility on Demand) AMoD jointly taking into account the decision making of drivers and the centralized coordinated fleet management?

Results: Development of a framework that models mobility-on-demand systems jointly considering various facets, including the drivers’ decision making and the decision making of the coordinated fleet, poses a challenging issue in modeling and simulation. The study [70] addressed this issue by proposing a comprehensive framework within an agent- and activity-based demand model, SimMobility. The model was integrated with a dynamic multi-modal network assignment model to accommodate behaviors of multiple on-demand services. A case study in Singapore based on the proposed framework was used to demonstrate the benefits of the proposed approach. The results indicated that AMoD produces a more efficient service. In addition, parking strategies and fleet sizes have significant influence on user satisfaction and performance.

(6)

Question: Is there a connection between life course and sustainable mobility for the millennial generation?

Results: The decreasing driver licensing and increasing use of sustainable modes in many countries suggest that the millennial generation might tend to more sustainable travel behaviors than previous generations. Whether millennials will continue to adopt sustainable transport modes as they age is an interesting research question. The study [53] provides an in-depth exploration of the connection between life course and mobility of millennials in Sydney, Melbourne and Canberra. The study [53] suggests that the difficulty in finding suitable housing near transit may push millennials into area where sustainable transport is impractical as they approach adult, get married or have children.

(7)

Question: What is the impact of the COVID-19 pandemic on public transportation?

Results: The COVID-19 pandemic had led to devastating economic and social disruption. The pandemic also had a significant influence on transport sector and great impact on transport patterns. The COVID-19 decreased travel, increased health and security concerns, and changed the travel patterns. The impact of COVID-19 pandemic on public transportation is an important research issue. The study [87] focused on the public transportation demands and patterns in Montevideo, Uruguay amid the COVID-19 pandemic. The study encompasses the analysis of trip reductions in Montevideo, the correlation between trip numbers and COVID-19 cases during the pandemic periods, the recovery of public transportation usage, and the correlation with socio-economic indicators. These results offer insight into the travel behavior of citizens throughout the COVID-19 pandemic and pave the way for policymakers and researchers to elaborate mobility strategies and policies.

(8)

Question: Is it possible to encourage people to make sustainable mobility choices and reduce car dependency/emissions and energy consumption through the development of a smartphone app?

Results: Although many cities and countries have been promoting sustainable transport, there lacks an innovative approach to guiding users/travelers to make sustainable mobility choices. In the study [54], a user centered application called GoEco! was proposed to persuade individuals to change their transport modes by exploiting eco-feedback, social comparison and automated mobility tracking. Approximately 150 voluntary users took part in the field test of GoEco! in Switzerland in a period of three months. The interviews with users and an online questionnaire provide insights into valuable future research directions to develop recommendations for similar persuasive apps.

(9)

Question: What is the impact of sustainable mobility strategies on reducing the use of private vehicles?

Results: Many countries and governments are introducing new mobility services and adapting the parking policies to promote sustainable mobility by reducing the use of private vehicles. An important research issue is to study the impact of sustainable mobility strategies. The paper [88] aims to study the impact of strategies on sustainable mobility for a metropolitan region in the Netherlands including Rotterdam and the Hague by employing activity based travel demand models (ABM). The case study finds that the impact due to introducing mobility hubs alone is limited. However, combining mobility hubs with sharing services has the potential to lead to significant reduction in the number of car trips. This study also finds that improving the public transport service with micromobility network, increasing parking costs and limiting parking capacity are helpful for reducing vehicle use and making mobility more sustainable.

(10)

Question: How to develop a framework for modeling traveler behaviors and incentivize travelers to select more sustainable travel options in on-demand mobility services?

Results: A framework for accounting the impacts of real-time on-demand mobility on traveler behaviors and modeling heterogeneous characteristics of consumers for integration with multi-modal dynamic simulators is essential for understanding traveler behaviors in response to on-demand mobility services. In [71], a framework considering sequential and inter-connected decision-making stages of on-demand service usage was proposed. The framework consists of a hybrid service subscription choice model, and logit mixture models for service access, menu product choice and opt-out choice. The framework was applied to a case to incentivize and encourage travelers to select sustainable travel options. The incentives and schedule delay perception were quantified by different population segments.

(11)

Questions: Is there a connection between transport mode choice and travel satisfaction? Is it possible to identify the connection between transport mode choice and travel satisfaction?

Results: Understanding the connection between travel satisfaction and mode choice is an essential research issue as it provides important insights into formulation of transport policies and prioritizing resources to achieve sustainable mobility patterns. The paper [55] focused on the study of the connection between mode choice and travel satisfaction by incorporating the model of human needs and travel satisfaction in the framework. The findings of this study include: (1) Car use satisfaction is positively related to transit use difficulties, and is negatively related to functional difficulties in car use and better cycling self-efficacy. (2) Bicycle satisfaction is related positively to cycling self-concepts and self-efficacy and is related negatively to car self-concepts. (3) Transit satisfaction is positively related to experiencing difficulties with other modes.

(12)

Questions: What are the factors that drive the adoption of ride-hailing and the associated travel characteristics? Are there mode substitution effects for ride-hailing?

Results: Studying the factors that drive the ride-hailing adoption and the associated travel characteristics is important for providing mobility-on-demand services. The paper [56] studies the factors that drive the ride-hailing adoption and mode substitution effects. The study finds that the factors influencing ride-hailing services adoption include socio-demographic factors, perceived benefits, ease of use, safety risks and car-dependent lifestyles. The study also finds that integration of other transport modes with ride-hailing is weak. This implies that ride-hailing is mostly used alone in door-to-door journeys.

(13)

Question: How to construct a simulation model to assess the potential impact of automated mobility-on-demand on urban mobility such as its disruptive effect on urban transportation for travelers currently using mass transit or private vehicles?

Results: Most studies on automated mobility-on-demand focus on the operational efficiency of the technology itself. The potential impact of automated mobility-on-demand on urban mobility such as its disruptive effect on urban transportation for travelers using mass transit or private vehicles need to be studied. The study [72] aimed to investigate the impact of automated mobility-on-demand on urban mobility. A flexible automated mobility-on-demand framework was developed for modeling/simulation based on SimMobility, a multi-modal agent-based urban simulation platform, which facilitates simulation of automated mobility-on-demand operations for different demand, such as mode choice, route choice and making trips. The study focused on the role of mass transit in the area where automated mobility-on-demand is widely available. The simulation results indicated that, to avoid congestion and maintain acceptable levels of urban transportation service, mass transit is irreplaceable.

(14)

Question: Do the characteristics and motives differ between business-to-consumer carsharing adopters and peer-to-peer carsharing adopters?

Results: Business-to-consumer carsharing and peer-to-peer carsharing are two business models for carsharing. Studying whether the characteristics and motives differ between B2C and P2P carsharing adopters is an important issue to promote carsharing. The study [57] focused on studying characteristics and motives of B2C carsharing adopters and P2P carsharing adopters in the Netherlands with different carsharing provider types to clarify their differences. The results of the study show that the characteristics of B2C and P2P carsharing adopters are very similar. The differences are in the frequency of using carsharing and public transport. The characteristics influencing a car owner to become an adopter of P2P carsharing as a provider were investigated. The study finds that car owners are much more likely to provide their car through an online platform if they already shared their car. The study suggests regulation to shape favorable conditions for connected multi-modal transportation systems rather than regulations for each carsharing business model.

(15)

Question: How to develop a model that incorporates various sustainability factors to provide decision support for sustainable household mobility choices in a multi-modal transport network?

Results: In the study [73], a Petri Net model considering various sustainability factors to capture the dynamics of multi-modal journey options in a complex transport network was proposed. The factors along with travel time and cost in the route optimization process were combined into impact score. The study demonstrated the capability of the model to direct users make more sustainable mobility choices through simulation. The model provides a tool that can be used by policymakers to assess the impact of a variety of sustainable transportation initiatives and infrastructure investments.

(16)

Question: How to design a scheme to reward travelers who have adopted sustainable transport modes?

Results: One way to promote sustainable mobility, alleviate congestion and reduce emissions is to reward travelers who have adopted sustainable transport modes. An interesting issue is to design a scheme to reward travelers who have adopted sustainable transport modes. In [58], a credit charge-cum-reward scheme was proposed. During a credit charge-cum-reward scheme period, a traveler may take transit to accumulate credits to minimize his/her travel costs. Alternatively, a traveler may choose to drive by consuming his/her credits. At the end of a credit charge-cum-reward scheme period, the government will settle the individual credit balance at a charging price for credit deficits or redemption for credit surpluses. An optimization problem was formulated based on consideration of the value of time and characterization of the periodic mode usage equilibrium. The authors defined a credit charge-cum-reward scheme to be Pareto improving if each traveler’s periodic travel cost is reduced under the scheme. The authors defined a credit charge-cum-reward scheme to be revenue neutral if the government gains zero revenue under the scheme. The study proves that the proposed credit charge-cum-reward scheme is Pareto improving and revenue neutral.

(17)

Question: How to determine and analyze the patterns of the modal accessibility gap in cities?

Results: There is a link between sustainability of cities and accessibility of different transport modes. Modal accessibility gap (MAG) refers to the spatial accessibility of different transport modes. Measuring MAG is one approach to assessing the degree of low-energy transport modes in cities. Measuring the MAG is an important issue in cities with multi-modal transportation networks. The study [108] measures the MAG in multi-modal transportation networks across the metropolitan area of Nanjing, China taking into account the factors of travel modes, traffic conditions, the location and attractiveness of the destination. The results indicate that there are substantial differences between public and private transport for a “door-to-door” trip in a multi-modal urban network environment. The accessibility of cars is always better than that of the transit and cycling modes. The smallest MAG between the transit mode and the car mode changes with traffic fluctuations and is distributed along metro and bus routes. MAG is influenced by conditions of road destiny, river obstacles, and the cross-river channel. The minimal MAG between the cycling mode and car mode is in central urban areas.

(18)

Question: How to assess the impact of home-to-work travel on sustainability?

Results: Assessing the impact of home-to-work travel on sustainability is an interesting and important issue as it requires application of tools and methodologies for the analysis. In [89], the authors proposed a methodology for assessing the sustainability of home-to-work travel with shuttle service in the ENEA Casaccia Research Centre based on post-cost–benefit assessment. The findings show that operating the home-to-work shuttle service leads to reductions of 97% of volatile organic compounds, 72% of particulate matter, and 60% of carbon dioxide. Furthermore, the analysis shows that the benefits for users and the community are estimated to be EUR 1.35 million per year.

(19)

Results: How to reconstruct a damaged urban system following earthquake to implement a more sustainable efficient urban transport system?

Results: Reconstructing a damaged urban system following earthquake relies on applying an appropriate method or principles. The research issue is not just to reconstruct the original urban but also to implement a more sustainable efficient urban transport system. An analysis of the influence of the earthquake on urban systems and mobility was carried out in [102]. An operating strategy integrating urban design with sustainable mobility strategies was proposed. The study shows that the post-quake reconstruction provides an opportunity to construct and implement a more sustainable efficient urban transport model through advanced technologies.

4.3. Potential Future Research Directions

The above review of the recent papers related to sustainability in the context of multi-modal mobility shows that the research issues or questions include: (i) factors influencing adoption of sustainable mobility behaviors, (ii) approaches to promoting sustainable mobility, (iii) factors with impact on transport modes and sustainability, (iv) modeling, simulation and planning in multi-modal transport networks, (v) urban policy design and implementation methods for sustainable urban mobility and (vi) performance evaluation in multi-modal transport systems. Several potential future research directions based on the review of the papers in the previous subsection are summarized in this subsection.

(i)

Factors influencing adoption of sustainable mobility behaviors

The paper [52] explored mobility motives of travelers to find the determinants for travelers to adopt sustainable mobility behaviors. The study [52] was conducted based on a sample of the Swiss population. Transferability of the results requires a further study. Consideration of implicit motives is an interesting extension to the study [52].

The study [53] explored the connection between life course and sustainability mobility behaviors of millennials in Sydney, Melbourne and Canberra. The study suggested policy interventions to support a sustainable lifestyle. However, the study did not cover rural Australia. Whether the results of this study hold true in other countries with different environments and constraints is an interesting research question. Further study may include the life course stages and typologies of mobility in survey research to determine their prevalence in the population.

Although the factors that drive the ride-hailing adoption and the associated mode substitution effects of ride-hailing were explored in [56], the study was limited. Use patterns of other household members such as children could be worth exploring further. In addition, safety perceptions and experiences of ride-hailing require further studies.

The model of human needs and travel satisfaction have been incorporated in the framework of the paper [55] to study the connection between mode choice and travel satisfaction. There are several future research directions related to the paper [55]: (1) studying the consistency of results across different groups in population with various travel purposes; (2) extending the study for leisure travel with joint travel; (3) studying the connection between mode choice and travel satisfaction for younger generations; (4) further study on per trip satisfaction of commuters (as the study [55] focuses on the overall satisfaction of commuters with their commuting choices overall rather than per trip).

(ii)

Approaches to promoting sustainable mobility

The study [54] explored the possibility to develop a smartphone app to enable and encourage users to make sustainable mobility choices, reduce car dependency/emissions energy consumption and conduct a field test based on the app. There are several future research directions stemming from the study [54] for further improvement: (1) extend the application with automatic travel routes and transport modes detection and a multi-modal travel planning system, to support users’ daily travel needs; (2) provide goal setting function, feedback, rewards at the individual level and at the community level or punishment and (3) exploit the power of social interactions.

Understanding the characteristics and motives of different types of carsharing adopters can help promote sharing based sustainable mobility. The differences between business-to-consumer carsharing adopters and peer-to-peer carsharing adopters in terms of characteristics and motives were studied in [57]. There are several future research directions following the study [57]: (1) to study the measures to convince people to share their cars and play the role of users and providers, (2) to study the potential groups that may benefit from carsharing but are still not using it much yet and (3) to extend the study for realistic datasets with other shared mobility modes and connection to public transport in different countries.

Rewarding schemes are one approach to promoting sustainable mobility. The study [58] focused on rewarding schemes for the commuting problem with a single origin–destination pair. Extension to multiple origin–destination pairs requires a further study. Analysis for travelers with non-recurrent and stochastic trips and other choices of travelers such as car ownership and vehicle types poses several challenging research issues needed to be studied in the future.

(iii)

Factors with impact on transport modes and sustainability

The study [87] focused on the impact of the COVID-19 pandemic on the public transportation system based on the case of Montevideo, Uruguay in terms of trip reductions, the correlation between trip numbers and COVID-19 cases, the recovery of public transportation usage, and the correlation with socio-economic indicators. The study sparks two future research directions. One future research direction is to broaden the analysis to encompass the factors such as age and tariff type. Another direction is to conduct a comprehensive analysis of the post-COVID-19 situation.

The study [88] assessed the impact of sustainable mobility strategies on reducing the use of private vehicles by employing activity based travel demand models. The authors of the paper [88] suggested several future research directions: (1) studying the effects of different model parameters in all dimensions; (2) improving the utility functions underlying the model; (3) improving the study based on more realistic assumption such as the fleet sizes of shared vehicles and (4) incorporating a network assignment model into the current travel demand model.

The study [89] assessed the impact of home-to-work travel with shuttle service on sustainability. Although the study [89] was limited to a single collective shuttle service, the approach is flexible and it has the potential to be applied to analyze other transport services. Inclusion of additional services such as carpooling and ridesharing requires further study. Another future research direction is to expand the current analysis to encompass the emerging trends of transport modes and transportation behaviors.

(iv)

Modeling, simulation and planning in multi-modal transport networks

The work [69] focused on the implementation of modality switch points within the road network to allow travelers to switch from one transport mode to the next to meet the requirements of versatile, individual, and spontaneous multi-modal travel. Although the core functionalities and algorithms of the model proposed in [69] are transferable to other urban settings, availability of data required to apply the model poses a challenge. A further study on the policy required to help meeting the requirements of data availability is needed to apply the proposed model and method.

Although the an agent- and activity-based demand model proposed in the work [70] can be applied to solve the challenging modeling and simulation issue of automated mobility-on-demand systems, there are two future research directions to improve the work: (1) quantifying the cost implications of automated mobility-on-demand system implementation strategies and comparing their respective benefits and (2) studying the driver behavior under the proposed framework.

In [71], a framework considering sequential and inter-connected decision-making stages of on-demand service usage was proposed. The framework consists of a hybrid service subscription choice model, and logit mixture models for service access, menu product choice and opt-out choice to model traveler behaviors in response to on-demand mobility services was developed. There are several future research directions following the study [71]: (1) investigating how other factors such as the ease of access to information influences the service access action; (2) collecting revealed preferences data for the framework; (3) extending to incorporate the en route opt-out behavior; (4) integrating the model into an ABM simulator for system-wide optimization.

The impact of automated mobility-on-demand on urban mobility was investigated in the study [72] based on a multi-modal agent-based urban simulation platform. The study [72] raises several research issues: (1) studying how to apply automated mobility-on-demand as a complementary service to mass transit, (2) optimization of fleet size and algorithms for re-balancing and ridesharing in automated mobility-on-demand, (3) studying policy implementations in automated mobility-on-demand, such as subsidizing automated mobility-on-demand services for connection to mass transit, (4) studying the effects of traffic control/management schemes such as transit signal priority or high-occupancy vehicle lanes and (5) redesign of mass transit services to work effectively with the availability of automated mobility-on-demand.

In the study [73], a Petri Net model considering various sustainability factors was developed to provide decision support for sustainable household mobility choices in a multi-modal transport network. There are several future research directions following the study [73]: (1) developing hybrid models by combining the PNs with the predictive power of machine learning algorithms to enhance the model by predictive and adaptive capabilities of AI, (2) development of advanced stochastic frameworks for handling uncertainty and stochasticity in transportation systems based on stochastic PN models, (3) scalable hierarchical or modular PN models for large scale transportation networks and (4) application of PNs to model and optimize emerging complex multi-modal transportation systems technologies and paradigms.

(v)

Urban policy design and implementation methods for sustainable urban mobility

The issues under this category are to develop an innovative and collaborative decision-making framework to support integrated sustainable urban policy design for sustainable urban mobility. In the study [101], a multi-level decision-making framework was proposed to support integrated sustainable urban policy design for sustainable urban mobility and applied to the metropolitan region of Rhine-Ruhr for verification. Following the study [101], there are several future research directions: (1) comparison of the Rhine-Ruhr region with other megacities to build a future international benchmark on urban networks to co-produce and translate relevant knowledge across the continents and (2) systematically integrating emerging business models in mobility and energy sectors into a policy-making framework through advanced digital technologies.

The issue to reconstruct a damaged urban system after earthquake to implement a more efficient urban sustainable transport system was addressed in the work [102]. The authors of [102] suggested several future research directions: (1) identification of the elements for integrating design models with sustainable networks, (2) analysis of key performance indices, and (3) investigation of policies for developing public transport in cities affected by disasters.

(vi)

Performance evaluation in multi-modal transport systems

The study [107] probed into the problem about how to estimate long-term travel delay in multi-modal systems with different transport modes. As the study was based on the Shenzhen city, China, the authors of the study [107] suggested a future research direction to apply the proposed method to other cities to verify transferability of the results.

The modal accessibility gap (MAG) refers to the spatial accessibility of different transport modes. The study [108] focused on measuring the modal accessibility gap (MAG) to assess the degree of low-energy travel mode in the city. The study [108] suggested several future research directions: (1) applying the proposed model to different population groups, (2) studying other factors influencing competition for public services among population groups, (3) studying the effects of the diversified distance decay coefficient or even function for different modes and (4) studying the effects of resolution of the raster cell scale.

5. Review of Studies Related to Shared Mobility or Ridesharing in the Context of Multi-Modal Mobility

In this section, we will present the details to find the papers relevant to shared mobility or ridesharing in the context of multi-modal mobility, review them and identify potential future research directions.

5.1. Summary of Studies

In Figure 3, the keyword “share” is detected. As “sharing” and “shared” are commonly used as alternatives to the verb “share” in the literature related to mobility, we used two query strings AB = ((“shared mobility” AND “multi-modal”)) and AB = ((“ridesharing” AND “multi-modal”)) to search the papers for the third category related to shared mobility and ridesharing in the context of multi-modal mobility. The sources of the literature were obtained based on the Web of Science database. The search resulted in 13 papers. After conducting content analysis for the papers obtained by searching the Web of Science Core Collection database with the above query strings, there are ten papers in this category. Figure 6 shows the distribution of these papers published over the years. The figure shows that the studies on shared mobility and ridesharing in the context of multi-modal mobility were continuous over the years. The review of these studies later in this section indicates that the studies on shared mobility and ridesharing in the context of multi-modal mobility is expected to continue in the upcoming years due to several unaddressed research issues.

The research questions addressed in each paper found by the refined search of this section are summarized in

Table 6. Summary of research questions addressed in the papers published between 2019 and 2022 in the Web of Science Core Collection with keywords “shared mobility/ridesharing” and “multi-modal” in the abstract.

Paper	Research Questions
[74]	How to solve the joint travel problem to find the optimal joint path to transport a set of travelers, where a joint path involves multiple paths of individuals associated with multiple origins and destinations?
[75]	How to develop methods or perform experiments to evaluate the operations of multi-modal multi-provider transport systems with different pricing schemes?
[103]	What are the requirements, challenges and future direction for the whole system design of intelligent shared mobility systems?
[90]	Is there a connection between ride-hailing/ridesharing adoption and transit service adoption?
[91]	How to develop a simulation framework for transportation planners to assess the impact of the emerging mobility modes with shared electric autonomous vehicles (AVs) on a city?
[104]	What are the types, variants, characteristics of shared mobility problems, the related solution approaches, the current trends and future research directions?
[76]	How to develop a decision model for travelers to choose routes and decide travel modes to minimize travelers’ generalized travel cost in a multi-modal transportation system with ridesharing and public transit?
[77]	How to simulate and optimize multi-modal transport systems with P2P ridesharing, transit, city bike-sharing and walking?
[78]	Can ridesharing help reduce traffic congestion in a multi-modal transport system ? If ridesharing cannot necessarily reduce traffic congestion, is it possible to reduce traffic congestion by implementing pricing strategies along with ridesharing?
[79]	How to optimally allocate and deploy designated autonomous vehicle (AV) lanes for shared autonomous vehicles (SAVs) considering flows of regular vehicles?

.

5.2. Review of Studies

The increasing uptake of shared mobility/ridesharing in recent years is reflected in the strong growth of shared mobility market. According to the reports of [126], the global shared mobility market size is expected to grow at a compound annual growth rate (CAGR) of 15.8% from 2023 to 2030. According to the reports of [127], the global shared mobility market size exhibits a compound annual growth rate (CAGR) of 12.00% during the forecast period (2024–2032). Despite the differences between the forecast of the compound annual growth rates, both reports project the strong growth of the global shared mobility in the upcoming years. The growing trend of the global shared mobility market in recent years has attracted a lot of attention from key stakeholders, including transportation planners, service providers, researchers, practitioners and travelers.

The emerging shared mobility and ridesharing transport modes have raised several interesting research questions for relevant stakeholders and posed challenges in their application. These research issues include requirements of shared mobility systems, travel planning, clarification of the connection between shared mobility and other transport modes, investigation of the impacts brought by shared mobility, decision models to determine the routes and travel modes to optimize performance or minimize travelers’ travel cost in multi-modal transport system with shared mobility/ridesharing and allocation of resources to meet the travel demand. Some of the research questions related to the research issues mentioned above along with the studies that respond to these questions are as follows.

(1)

Question: How to solve the joint travel problem to find the optimal joint path to transport a set of travelers, where a joint path involves multiple paths of individuals associated with multiple origins and destinations?

Results: Solving the problem to transport a set of travelers is an essential issue in ridesharing systems. The problem to transport a set of travelers in a travel group is called a joint travel problem. This problem was addressed in [74]. The issue to solve the joint travel problem is to find the optimal joint path for a travel group, where a joint path involves multiple paths of individuals associated with multiple origins and destinations. A space–time multi-state super-network framework was proposed to address the joint travel problem. The optimal joint paths are found by developing stage-wise recursive formulations. Analysis indicates that the run-time complexity for the problem does not depends on the number of meeting/departing points. The study indicates that the joint travel problem is a variant of the Steiner tree problem and is NP-hard.

(2)

Question: How to develop methods or perform experiments to evaluate the operations of multi-modal multi-provider transport systems with different pricing schemes?

Results: In cities with multi-modal transport networks, there are multiple transport modes. Different transport modes may be offered by different service providers with different pricing schemes. Such a multi-modal multi-provider market relies on developing methods or performing experiments to evaluate the operation of transport systems. A generalized market equilibrium model based on Karush–Kuhn–Tucker (KKT) optimality conditions of the optimization problem formulation was proposed in the study [75] to assess the operation of transport systems with private vehicles, walking, public transport, ridesourcing and ridesharing. Experiments were performed by applying the proposed model to an abstract network in Sydney to demonstrate its effectiveness in handling responses of market players to electric vehicles, changes in travel demand, greenhouse gas emissions limit, and barriers of ridesharing.

(3)

Question: What are the requirements, challenges and solutions for the whole system design of intelligent shared mobility systems?

Results: Although the increasing number of papers on Shared Mobility Systems leads to several survey papers that summarize how a particular challenge is addressed individually, there lacks a survey from the “whole system design” point of view. A survey of flexible shared mobility systems supporting decision making for inter-operable multi-modal transport services is needed. The paper [103] categorizes the inter-connected challenges in different transportation modes of shared mobility systems and reviews how the resulting solutions address these challenges in the system.

(4)

Question: Is there a connection between ride-hailing/ridesharing adoption and transit service adoption?

Results: Clarifying whether there is a connection between ride-hailing and transit service adoption is helpful for understanding the inter-modal complementarities and planning of multi-modal transport systems. Therefore, studying the connection between ride-hailing and transit service adoption is an important issue. The study [90] empirically identified the relationship between transit service and ride-hailing adoption in a fast-growing region in the United States. The study found that the surge periods for these two transport modes were not simultaneous temporally. The study also found that ridesharing complements transit adoption.

(5)

Question: How to develop a simulation framework for transportation planners to assess the impact of the emerging mobility modes with shared electric autonomous vehicles (AVs) on a city?

Results: Studying the potential impact of the emerging mobility modes with shared electric autonomous vehicles (AVs) on a city daily life is an important issue for transportation planners. In [91], an agent-based simulation framework was proposed traffic planners to explore the cooperative integration of AVs using a decentralized control approach and assess the impact of the emerging mobility modes through simulation. A prototype was implemented and validated based on the data of Trento.

(6)

Question: What are the types, variants, characteristics of shared mobility problems, the related solution approaches, the current trends and future research directions?

Results: The emerging shared mobility services not only receive widespread adoption by cities to achieve the sustainability goals and high utilization by users to benefit from low cost and accessible journey in daily life but also attract the attention of numerous researchers, practitioners and related stakeholders. This leads to a large number of studies and papers on shared mobility problems. A systematic review of shared mobility problems is required to provide insights into the current trends and future research directions. A systematic review of the literature on shared mobility problems according to problem types, variants, characteristics, and solution methods is provided in the paper [104]. The findings are that the main issue in the shared mobility literature is the total cost minimization for static and dynamic cases. Shared mobility problems considering time window and multi-objective characteristics are popular among the research community. The most widely adopted solution approaches to shared mobility problems are heuristic and metaheuristic approaches.

(7)

Question: How to develop a decision model for travelers to choose routes and decide travel modes to minimize travelers’ generalized travel cost in a multi-modal transportation system with ridesharing and public transit?

Results: In a multi-modal transportation system with ridesharing and public transit, travelers must choose routes and decide travel modes. A traveler may play the role of a solo driver, ridesharing driver, ridesharing passenger, or transit passenger. A decision model must be developed for this problem. In [76], a multi-modal route choice model with ridesharing and public transit was developed based on single-origin–destination (OD) pair network and was expressed as an equivalent complementarity problem. The results in equilibrium indicate that the travelers’ route and mode choice behavior is significantly influenced by rewards of passengers’ and toll charge of solo drivers. The results also show that increasing passengers’ rewards can attract more travelers to adopt environmentally friendly transport modes.

(8)

Question: How to simulate and optimize multi-modal transport systems with P2P ridesharing, transit, city bike-sharing and walking?

Results: Although P2P ridesharing, transit, city bike-sharing and walking have been implemented separately in cities, there lacks a study on the integration of these heterogeneous transport modes and effectiveness of the resulting integrated system. A study based on simulation of multi-modal transport systems with P2P ridesharing, transit, city bike-sharing and walking is needed. In [77], a scheme to offer travel alternatives across multiple transport modes, including P2P ridesharing, transit, city bike-sharing and walking, were developed. By applying the proposed method to the case of the multi-modal system of Los Angeles, a simulation enhanced with optimization was conducted. The study found that properly designed ride ridesharing and bike-sharing, when integrated into the transit system, can form a multi-modal network that expands the coverage of public transit.

(9)

Questions: Can ridesharing help reduce traffic congestion in a multi-modal transport system ? If ridesharing cannot necessarily reduce traffic congestion, is it possible to reduce traffic congestion by implementing pricing strategies along with ridesharing?

Results: For a multi-modal transport system with ridesharing, an interesting research issue is to study whether ridesharing can help reduce traffic congestion. If ridesharing cannot necessarily reduce traffic congestion, is it possible to reduce traffic congestion by implementing pricing strategies along with ridesharing? In [78], different pricing strategies were proposed and compared for a multi-modal network with ridesharing services. A model was developed based on an aggregate traffic representation, the Macroscopic Fundamental Diagram (MFD), to study how time-dependent travelers’ choices and traffic conditions evolve from day to day. Two different congestion pricing schemes were proposed and compared to attempt to reduce traffic congestion and improve efficiency. One scheme prices all vehicles including both solo-driving and ridesharing vehicles, while the other scheme prices the solo-driving vehicles only to encourage ridesharing. The study indicates that ridesharing may not necessarily reduce traffic congestion. The results show that pricing solo-driving vehicles only might encourage ridesharing but it is less effective in reducing the overall congestion in comparison with pricing both solo-driving and ridesharing vehicles.

(10)

Question: How to optimally allocate and deploy designated autonomous vehicle (AV) lanes for shared autonomous vehicles (SAVs) considering flows of regular vehicles?

Results: Shared autonomous vehicles (AVs) have the potential to become a popular sustainable transport mode. Deploying designated autonomous vehicle (AV) lanes for shared autonomous vehicles (SAVs) considering flows of regular vehicles is one approach to enhancing operational efficiency. Development of a solution to allocate AV lane capacity is needed to provide effective support for managers to allocate dedicated AV lane. In [79], the dedicated AV lane optimization allocation problem considering mixed SAVs and regular vehicles flows was studied. An enumeration algorithm was proposed to develop lane allocation strategies for discrete lane settings. The findings of this study was that the deployment of dedicated AV lanes should jointly consider the penetration rate of SAVs and the management objectives. In addition, the authors performed sensitivity analysis taking into account the value of travel time for SAVs and the capacity of a dedicated AV lane.

5.3. Potential Future Research Directions

The above review of the recent papers related to shared mobility and ridesharing within the context of multi-modal mobility highlights several research issues and questions requiring further investigation. These research issues or questions include (i) modeling, simulation and planning in multi-modal transport networks, (ii) leveraging the benefits of shared mobility to enhance multi-modal transport and (iii) studying the relation between different transport modes and the impacts of transport modes. These potential future research directions are summarized in this subsection.

(i)

Modeling, simulation and planning in multi-modal transport networks

Although the problem to transport a set of travelers in a travel group called a joint travel problem has been addressed in [74], there are several interesting directions for future research related to [74], including (1) studying the use of personalized preferences and choice making mechanisms to achieve accurate synchronizations among the individuals, (2) developing location selection models taking into account space–time constraints, coupling constraints and vehicle capacity constraints to reduce the run-time complexities, (3) extending the study to deal with the joint travel problem with multi-modal, multi-person and multi-activity, (4) development of techniques to speed up the solution finding process and (5) extending the study to deal with the joint travel problem under uncertainty.

Evaluation of the operation of transport systems with different transport modes, including private cars, walking, public transport, ridesourcing, ridesharing-as-driver, and ridesharing-as-rider, has been studied in [75]. However, the model used in the study [75] is not general enough to deal with situations in the real-world environment. Including other players in the formulation, considering energy supply market players of electric vehicles in the model and inclusion of stochastic aspects in the model are interesting research directions. In addition, the ridesourcing service is defined over the links in the proposed model. Extension to deal with origin–destination requirement is required to solve a problem in the real world.

In [76], a multi-modal route choice model with ridesharing and public transit was developed based on single-origin–destination pair network and was expressed as an equivalent complementarity problem. There are several future research directions following the study [76]: (1) consideration of departure time choice, (2) extension of the proposed model to multi-user classes with different values of time or ridesharing preferences to study the influence on the equilibrium flow patterns and (3) studying the design of traffic policies and their effects on the ridesharing behavior.

Although the study [77] showed that the scheme to provide travel alternatives across multiple modes, such as P2P ridesharing, transit, city bike-sharing and walking, can expand the coverage of public transit, it only considers travel demand from personal vehicles. Consideration of transit demand is an interesting issue to study the potential improvements and mode shift. Other behavior models, e.g., random utility models could be studied and implemented in the future research. To study the effects of pricing, transfer and waiting time on riders’ choice behavior, sensitivity analysis is required.

The work [78] developed based on an aggregate traffic representation to study how time-dependent travelers’ choices and traffic conditions evolve from day to day was limited. There are several directions for further research: (1) extension to account for more complex networks and traffic models, (2) studying the feasibility to replace the aggregate traffic model with a detailed link-node model, or cell-transmission model, or link-transmission model, (3) extension to adopt alternative multi-modal MFD to incorporate flow interaction between car traffic and transit vehicles and (4) extension to stochastic MFD.

For the dedicated AV lane optimization allocation problem considering mixed SAVs and regular vehicles flows in [79], several directions are worth exploring: (1) extension to incorporate time-varying congestion tolling strategy in the dedicated AV lane optimization allocation problem, (2) studying the effectiveness of allowing commuters to pay for riding on the AV lane, (3) combining the actual traffic system to study lane management on a larger scale of road networks, (4) considering the characteristics of user heterogeneity, (5) studying travel patterns during the evening rush hour and (6) studying the effectiveness of temporarily converting dedicated bus lanes into general-purpose lanes in case the penetration rate of AVs is low.

(ii)

Leveraging the benefits of shared mobility to enhance multi-modal transport

The paper [103] provides a survey from the “whole system design” point of view in the context of shared mobility systems. The paper suggests several future research directions for achieving “whole system design”, including (1) design a system as a whole, (2) design a system with autonomy at a higher level, (3) towards an integrating plug-and-play architecture, (4) towards a multi-hop and multi-mode shared mobility and (5) design a system with robustness to face unexpected behaviors or events and autonomous to reduce inter-dependencies and ensure better resilience.

A systematic review of the literature on shared mobility problems based on the problem types, variants, characteristics, and solution approaches is provided in the paper [104]. There are several directions related to [104] for further study: (1) adoption of hybrid algorithms, (2) implementation of shared mobility problems into Mobility as a Service, (3) adoption of electric vehicles in shared mobility problems and (4) combined people and freight transportation service with ridesharing in multi-tier and multi-modal transport networks.

(iii)

Studying the relation between different transport modes and the impacts of transport modes

It was unclear whether ride-hailing/ridesharing compete or complement transit service. The study [90] clarifies the relationships between ride-hailing/ridesharing adoption. The study found that the surge periods for these two transport modes were not simultaneous temporally and ridesharing services complement transit adoption. The study [90] suggests that more studies are needed to inform policy on the contributions of more on-demand multi-modal transport options and negative externalities on increasing vehicle mile traveled, traffic congestion, bus scheduling delays, and urban air quality.

For the study [91] of the impact of the emerging mobility modes with shared electric autonomous vehicles (AVs) on a city, there are two potential future research directions: (1) the implementation of adaptation strategies on the framework to handle adaptation needs, such as dynamic rescheduling of the available AVs and re-grouping of passengers of a faulty car, among the closest AVs around and (2) adding explanation capabilities about the decision making and impact analysis information to the framework.

6. Review of Studies Related to Micromobility in the Context of Multi-Modal Mobility

In this section, we will present the details to find the papers relevant to micromobility in the context of multi-modal mobility, review them and identify potential future research directions.

6.1. Summary of Studies

In Figure 3, the keyword “bike” is detected. As “bike” is one type of vehicles in micromobility, it can be selected as a candidate keyword to search for related papers. As “scooter” is another type of vehicles in micromobility, it can be selected as a candidate keyword to search for related papers. We used two query strings AB = ((“bike” AND “transport” AND “multi-modal”)), AB = ((“micromobility” AND “multi-modal”)) and AB = ((“scooter” AND “transport” AND “multi-modal”)) to search for the papers for the fourth category related to micromobility in the context of multi-modal mobility. The sources of the literature were obtained based on the Web of Science database. The search resulted in 17 papers. Although query strings were used to search for papers potentially relevant to the combination of keywords for the review, many papers containing the keywords in the abstract did not focus on research issues related to those keywords. If a paper was not directly related to the research issues associated with the keywords after content analysis, it was excluded from this review. After screening the papers found by searching the Web of Science database, 11 papers were obtained.

Figure 7 shows the distribution of these papers published over the years. Although there was a break in publication in 2022, the sustained growth in the number of publications in 2023 and 2024 indicates that the number of related publications has the potential to grow in the coming years, helping to fill the gap in the literature.

The research questions addressed in each paper found by the refined search of this section are summarized in

Table 7. Summary of research questions addressed in the papers published between 2017 and 2024 in the Web of Science Core Collection with keywords “bike/micromobility/scooter”, “transport” and “multi-modal” in the abstract.

Paper	Research Questions
[92]	What are the impacts of bike-sharing on travelers’ usage of other transport modes in a multi-modal transportation system?
[59]	How does multi-modal travel improve accessibility of tourist attractions?
[93]	How to assess the mode substitution effects of dockless bike-sharing systems and its determinants?
[94]	How to assess the interactions between different modes of transport and the influence of one transport mode on the other one from the perspectives of time and region attribute?
[95]	What are the determinants of station-based round-trip bike sharing demand?
[60]	How to model public transport users’ propensity to shift to the bicycle to access bus stops, train and metro stations?
[61]	How to analyze bike sharing travel patterns based on bike sharing trip data?
[96]	What are the differences of micromobility modes in terms of the characteristics, complementation or competition between them?
[62]	How to optimize the locations of micromobility stations/service area using geospatial data?
[63]	How to assess the effectiveness of integrating shared e-scooters as the feeder to public transit in cities?
[64]	How governance stakeholders might shape sustainability transitions and technology uptake of e-bike?

.

6.2. Review of Studies

Despite safety concerns [128], safety hazards [129] and anti-social behavior by riders such as theft, vandalism, carelessly leaving bikes or scooters in inconsiderate locations, obstructing pedestrians/disabled persons, the world has witnessed strong growth in the micromobility market since its inception due to its potential sustainability benefits [130]. According to the reports of [131], the global micromobility market is expected to grow from USD 175 billion in 2022 to USD 360 billion by 2030. The report of [132] projects that the compound annual growth rate (CAGR) is 13.00% during the forecast period (2024–2030) and the report of [133] estimates that the compound annual growth rate (CAGR) is 12.5% during the forecast period (2022–2027). The emerging micromobility transport mode has created several interesting research issues/questions that pose challenges in its application. When a new transport mode appears in the existing transportation system, it may raise several new issues not addressed before. These new issues include the assessment of the impact on other ones caused by the new transport mode, the potential benefits brought by the new transport mode, effective approaches to integrating the new transport mode in a multi-modal network to attain the potential benefits, the determinants of adopting the new transport mode and the development of methods for the optimization of planning/design of the infrastructure in terms of cost, accessibility, efficiency, performance and sustainability, etc. Some of the research questions related to the new issues mentioned above have been studied in recent years. These research questions along with the studies that respond to them are as follows.

(1)

Question: What are the impacts of bike-sharing on travelers’ usage of other transport modes in a multi-modal transportation system?

Results: The study [92] focused on developing a spatial agent-based model to assess the impacts of bike-sharing on travelers’ usage of other transport modes through simulation. The results showed that providing free bike-sharing to connect the transit system can result in savings of 1.5 million US dollars in transportation damage cost per year, and preventing 22 premature deaths per year due to mode shift to cycling and walking for the city under study. However, the impacts of bike-sharing on the use of motorcycles was limited. The proposed model provides policy makers with a tool to improve the sustainability of a multi-modal transportation system with bike-sharing.

(2)

Question: How does multi-modal travel improve accessibility of tourist attractions?

Results: Due to the assumption of uniform access for each transport mode within the catchment area, applying the two-step floating catchment area (2SFCA) method results in inaccuracy in measuring the accessibility of tourist attractions. Development of a more effective method to measure the accessibility of tourist attractions accurately is urgent. A multi-modal network incorporating a super-network with the two-step floating catchment area (MMN-2SFCA) method was proposed in [59] to model transfer behavior, travel mode choice, and travel demand estimation in multi-modal travel by multi-source big data including subway data and bike-sharing data to improve the accuracy in measurement of tourist attraction accessibility. The effectiveness of the proposed method was demonstrated by the results of the case study in Suzhou.

(3)

Question: How to assess the mode substitution effects of dockless bike-sharing systems and its determinants?

Results: An important issue of dock-less bike sharing is to study the mode substitution with respect to other transport modes at the trip level to assess their societal and environmental impact and study the complex effects of built environment factors on the mode substitution patterns of dockless bike sharing by applying machine learning to multiple data sources. The results presented in [93] showed that the probabilities for dockless bike sharing to replace other transport modes can be quantified by the proposed method at the trip level and were heterogeneous depending on the trips in urban contexts. The average rates that bike-sharing may substitute bus, metro, walking and ride-hailing were projected to be 0.356, 0.116, 0.347 and 0.181, respectively, in Shanghai. The rate for dockless bike sharing to substitute a certain transport mode can be linked to built environment factors such as presence of transit systems.

(4)

Question: How to assess the interactions between different modes of transport and the influence of one transport mode on the other one from the perspectives of time and region attribute?

Results: Exiting studies on transport demand forecasting primarily focus on a single transport mode based on the associated historical data. There lacked a study on co-evolution between different transport modes, e.g., the interactions between different modes of transport and influence of one transport mode on the other one from the perspectives of time and region attribute. In [94], the co-evolution between taxi demand and shared-bike demand was studied. There are two challenges for discovering the co-evolving patterns of different transport modes: (1) diversity of the co-evolving correlation in terms of regions and time and (2) multi-modal data fusion. The authors overcame these challenges by developing a novel method called the coevolving spatial temporal neural network to learn a multi-view demand for each transport mode, extract the co-evolution pattern of the two modes and predict the demand for a specific transport mode. The proposed approach uses a multi-scale representation method, including fine-scale demand information and coarse-scale pattern information. The superiority of the proposed approach over the state-of-art methods was demonstrated by the results of extensive experiments.

(5)

Question: What are the determinants of station-based round-trip bike sharing demand?

Results: A train station-based round-trip bike sharing scheme, which was different from the existing one-way bike sharing schemes, was promoted by the Dutch train station operator to offer bike and train services. However, there lacked a study on the determinants of demand for round-trip bike sharing. The paper [95] identified potential temporal and weather-related determinants for station-based round-trip bike sharing rentals by applying multiple linear regression with an in-depth analysis. The authors compared the results with the findings of one-way bike sharing schemes. The results show that the most important determinants for hourly rentals in the station-based round-trip bike sharing scheme is the number of departing train travelers, followed by temporal and weather-related determinants. In addition, the correlation between the determinants and the hourly demand varies with stations and depends on the underlying demand patterns.

(6)

Question: How to model public transport users’ propensity to shift to the bicycle to access bus stops, train and metro stations?

Results: Although the integration of cycling and public transport in developed countries has been studied, the issue to integrate cycling with other main modes such as bus and metro in developing countries is less explored. The goal of the study [60] is to study the propensity for public transport users to shift to the bicycle to access bus stops, train and metro stations in Rio de Janeiro. This study used two binary logit models to predict the factors affecting the propensity to adopt a bicycle to access public transport. The findings are that socio-economic characteristics, barriers and motivators are important factors for users to adopt bike and ride. Personal constraints, distance too close to public transport stops, parking conditions and public safety are the barriers. Changing home location, owning a bicycle, implementing cycle ways and improving parking conditions are motivators. The study suggests formulation of policy to increase bicycle owner-ship and improve cycling infrastructure.

(7)

Question: How to analyze bike sharing travel patterns based on bike sharing trip data?

Results: Bike sharing travel patterns can provide insights about the use, design and policy decisions of bike sharing and the associated multi-modal transportation systems. An important research issue is to analyze bike sharing travel patterns based on bike sharing trip data. The study [61] analyzed the distributions of trip distance and trip duration for bike sharing trips for commuting and touristic purposes based on bike sharing trip data from eight cities in the United States. The results show that, for larger bike sharing systems, both the trip distance and duration follows a lognormal distribution. For smaller systems, the distribution varies among Weibull, gamma, and lognormal. For the case of long trips, the analysis shows that the trip distance and duration follows a power law decay in the larger systems.

(8)

Question: What are the differences of micromobility modes in terms of the characteristics, complementation or competition between them?

Results: The combination of shared bikes, shared e-scooters, and public transit enables an efficient and flexible multi-modal transport service. However, differences of micromobility modes were not explored in the literature. An investigation was required to study the characteristics of different modes of micromobility as well as the complementation or competition between them. The study [96] was based on the case of the City of Austin with coexisting public transit, shared bikes, and shared e-scooters. The findings of this study indicate that public transit is closely related to commuting than shared micromobility modes. The results show that e-scooters are more likely to be used for leisure trips than shared bikes and public transit. In comparison with shared e-scooters, shared bikes are more likely to be used by commuters during peak hours. The results suggest that only one shared mode complements public transit at a given area.

(9)

Question: How to optimize the locations of micromobility stations/service area using geospatial data?

Results: Optimization of the locations of micromobility stations/service area poses several challenges for smaller urbanized areas with lower population densities due to fewer resources for system planning, operation, and maintenance. The development of an effective solution method is an important issue. A method based on rasterized geospatial data to capture bikeshare demand indicators was developed in [62] to optimize micromobility stations/service area locations. Priority of inputs with weights calculated by an analytic hierarchy process based on the results of a survey of transportation professionals was used to reflect the overall importance of different inputs. The proposed method identified hotspots for candidate stations/service area locations by combining different factors and creating a spatial index value. The candidate stations/service area locations were further analyzed to choose optimal locations based on the factor of budget and spacing between stations/service areas. The proposed method was applied to the bikeshare station siting case of Iowa City, Iowa, U.S. for validation. The results show that the introduction of optimal locations found improve multi-modal travel times and job accessibility in the target area.

(10)

Question: How to assess the effectiveness of integrating shared e-scooters as the feeder to public transit in cities?

Results: As a complementary transport mode to solve the first- and last-mile problem to access public transit, the integration of e-scooter sharing with public transit is an important issue. The study reported in [63] was based on vehicle transport data of 124 European cities to analyze the integration between shared e-scooters and public transit. The analysis was based on the integration ratio, which is the percentage of trips using e-scooter sharing as a feeder to public transit over all e-scooter sharing trips. The analysis shows that the integration ratios of e-scooter sharing vary from 5.59% to 51.40% with a mean value of 31.58% and a standard deviation of 8.47%. Analysis of the temporal patterns shows that an increase in the integration ratio for first-mile trips is associated with a decrease in the integration ratio for last mile in the time series. By applying a bottom–up hierarchical clustering method, the analysis of the study in [63] divides these cities into four clusters according to their temporal variations in the integration ratios. By applying a machine learning approach to analyze the impact of city-level factors on the integration ratio, it shows that the public transit stations density and the ratio of the young are positively associated with the integration ratio.

(11)

Question: How governance stakeholders might shape sustainability transitions and technology uptake of e-bike?

Results: Governance stakeholders play a pivotal role in taking action toward sustainable development through promoting enterprise transformation and industrial transformation. However, existing studies on the role of governance stakeholders in shaping e-bike technology uptake/rejection is limited. Understanding how governance stakeholders might shape sustainability transitions and technology uptake is an important issue. The study [64] provides insight into how governance stakeholders shape sustainability transitions and technology uptake by exploring their perspectives in depth. This study finds that e-bikes possess niche benefits for helping disrupt auto-centric socio-technical systems through strengthening first and last mile connectivity, multi-modal integration and shared mobility innovations. To benefit from e-bikes, distinction between e-bike models, clarity on usage and transition at the city scale are required.

6.3. Potential Future Research Directions

The above review of the recent papers related to bike, micromobility and scooter in the context of multi-modal mobility indicates that several research issues or questions have been studied. Several potential future research directions based on the review of the papers in the previous subsection are summarized in this subsection.

(i)

Impacts and substitution effects of micromobility

The study [92] focused on developing a spatial agent-based model to study the impacts of bike-sharing on travelers’ usage of other transport modes through simulation. There are several directions for future research: (1) introducing weather factor in the model and studying the effects of weather, (2) studying the leisure travel and (3) incorporating psychological factors such as comfort and perceptions of safety in the agent decision-making model.

Although the study [93] lacks consideration of the substitution of private cars by dockless bike sharing, the proposed analysis framework is general and can be used for analyzing the mode substitution of e-scooter and e-moped sharing. An interesting future research direction is to compare the mode substitution rates of dockless bike sharing and e-scooter sharing systems. Performing similar analysis for other cities with distinct urban structures/culture and comparing the results to identify potential new findings is an interesting research direction.

The study [94] focused on the co-evolution between taxi demand and shared-bike demand to understand the connection between these two transport modes. An interesting direction for further study is the investigation of the co-evolution patterns between other transport modes such as subway demand and taxi demand.

The paper [95] identified potential temporal and weather-related determinants for station-based round-trip bike sharing rentals by applying multiple linear regression and an in-depth analysis. Validity about whether the findings of the study hold during post-COVID-19 period requires further study. As the limitations of the proposed approach is due to multiple linear regression, conducting a similar study based on advanced methods to identify determinants is an interesting issue. Capturing other determinants of station-based round-trip bike sharing, e.g., events and service disruptions, is a potential future research direction.

The work [96] studied the characteristics of different modes of micromobility as well as the complementation or competition between them. The results of [96] were obtained without the information of the actual trip purpose of the different traffic modes. A future research direction is to incorporate a survey of trip purposes to derive more detailed results, including the usage between micromobility and geographical information as well as the reasons for using micromobility.

(ii)

Characteristics and adoption factors of micromobility in multi-modal transport networks

The analysis of the distributions of trip distance and trip duration of bike sharing travel patterns in the study [61] was based on bike sharing trip data with only trip origin, destination time and location. An extended study is required for the detailed trajectory data of the bike sharing trips. The study [61] focuses on the analysis of the bike sharing travel patterns for station-based bike sharing systems. Analysis of dockless bike sharing systems is an interesting future research direction.

The goal of the study [60] was to study the propensity for public transport users to shift to the bicycle to access bus stops, train and metro stations in Rio de Janeiro. There are several directions for future research:

(1)

An urgent research issue is to study the effectiveness of using bicycle to access different transport modes (metro, bus and train).

(2)

The results of the study were based on work/school trips. Whether the results will be the same for other trip purposes requires a further study.

(3)

Incorporating other psychological indicators to study the propensity for public transport users to shift to the bicycle to access public transport is an interesting research issue.

The study [59] modeled transfer behavior, travel mode choice, and travel demand estimation in multi-modal travels by multi-source big data to improve the accuracy in measurement of tourist attraction accessibility. An interesting issue is to plan multi-modal transport routes/itineraries to visit attractions based on the tourist attraction accessibility.

The study [63] analyzed the integration between shared e-scooters and public transit in European cities. There are several directions for future research:

(1)

The analysis of this study was based on the Euclidean distance between trip origin/destination and the associated nearby public transit stations. More accurate integration analysis based on the earliest departure time from origin and the latest arrival time to get to destination, public transit schedules and the points of interest around trip end-points is a potential future research direction.

(2)

Incorporation of missing potential influencing factors such as education level and income and studying their impacts on the user behavior and integration of e-scooter sharing with public transit are an interesting issue.

(3)

The study [63] focuses on the integration of e-scooter sharing with public transit modes. Consideration of both the complementary and substitution roles of e-scooter sharing to investigate their impacts on public transport is an important issue.

Although the problem to optimize micromobility stations/service area locations had been studied in [62], consideration of the existing public transit station locations and schedules in the optimization method is a challenging issue.

The study [64] provides insight into how governance stakeholders might shape sustainability transitions and technology uptake by exploring their perspectives in depth. There are several directions for further studies. Impact on certain sub-demographics (e.g., women, older and/or low-income demographics) due to adverse disruptions remains an important area for future research. Greater attention should be paid to the role of governance actors in the sustainability transition processes (e.g., implementation, monitoring and evaluation stages). Explore how governance actors operate within and across jurisdictions to coordinate and influence one another across networks to understand the diffusion of transition across jurisdictions.

7. Review of Studies Related to Integration in the Context of Multi-Modal Mobility

In this section, we will present the details to find the papers relevant to integration in the context of multi-modal mobility, review them and identify potential future research directions.

7.1. Summary of Studies

As MaaS relies on integration to provide seamless, high-quality door-to-door transport services for users, “integration” can be selected as a candidate keyword to search for related papers. We used the query strings AB = ((“MaaS” AND “integration”)) and AB = ((“MaaS” AND “pricing”)) to search for the papers for the fifth category related to integration issue and pricing issue in the context of multi-modal mobility. The sources of the literature were obtained based on the Web of Science database. The search resulted in 98 papers. Although query strings were used to search for papers potentially relevant to the combination of keywords for the review, many papers containing the keywords in the abstract did not focus on research issues related to the keywords. If a paper was not directly related to the research issues associated with the keywords after content analysis, it was excluded from this review. After screening the papers found by searching the Web of Science database, 21 papers were obtained. Figure 8 shows the distribution of these papers published over the years. Although the number of publications did not increase monotonically, publications continued to grow in 2024 to fill the gap in the existing literature.

The research questions addressed in each paper found by the refined search of this section are summarized in

Table 8. Summary of research questions addressed in the papers published between 2017 and 2024 in the Web of Science Core Collection with keywords “MaaS” and “integration/pricing” in the abstract.

Paper	Research Questions
[105]	Are consumers willing to pay for collaboration activities arising from institutionally facilitated interaction between operators?
[80]	How to model demand and plan for ride-shared mobility services running in integration with mass transit?
[81]	How to evaluate the integrated behavior of MaaS?
[115]	Can consumers benefit from bundling transportation services?
[109]	How to integrate freight transport into a Mobility-as-a-Service (MaaS) environment?
[110]	How to achieve operational, informational and transactional integration required for MaaS to deliver an exceptional user experience to rival the private cars? What factors affect how people make travel choices for adoption of any MaaS offer?
[111]	How to extend the concept of Mobility as a Service (MaaS) to fully integrate public transportation to make public transport an attractive alternative to private transport?
[106]	How to systematically support alliance formation in MaaS?
[112]	How can a public transport authority develop MaaS in rural areas through integrating a public transport service with carpooling?
[82]	How to simulate autonomous mobility-on-demand systems to evaluate its impact from the perspectives of travelers, operators and the city?
[113]	What are the characteristics of a MaaS service and how to compare different services, understand the potential effects of MaaS, and integrate the societal goals into MaaS services?
[83]	How to assess the effects of ridesourcing services on service providers and the transport system by simulation?
[114]	How to integrate one-to-many peer-to-peer ridesharing with public transit on a MaaS platform for morning commute?
[116]	What are the consumers’ intention to subscribe to MaaS, preferences for bundling in MaaS and willingness to pay for extra features of MaaS?
[117]	How to evaluate added values of MaaS bundles for users with heterogeneous subscription willingness?
[118]	How do different MaaS bundling and pricing schemes contribute to sustainable transportation?
[119]	How to allocate costs and determine prices offered by operators according to user route choices and operator service choices in MaaS systems?
[120]	How to solve the vehicle dispatching service pricing and demand problem in MaaS?
[121]	How to balance sustainability and profitability in electric Mobility-as-a-Service ecosystems by providing carbon emissions reduction incentives?
[122]	Can a MaaS system ensure that travelers and service operators will be better off in terms of welfare while maintain profitability of the platform? Does there exist a pricing scheme that can ensure stability condition under which operators and travelers of a MaaS system are willing to participate?
[123]	How to develop an insurance product that can cover the policyholders throughout their entire trips with different transport modes? How to incentivize sustainable multi-modal transport through an innovative risk assessment and transfer scheme?

.

7.2. Review of Studies

MaaS aims to provide more sustainable, flexible and efficient door-to-door transportation services based on various forms of transport modes. To achieve this goal, it relies on integration of various transport modes at different levels of the multi-modal transportation systems to facilitate the processes from planning travel, booking tickets to payment for travelers. Also, it relies on the collaboration of various stakeholders such as the service providers of different transport modes and the transportation authority to implement MaaS systems. The requirements to integrate different transport modes, support travel planning/ticketing processes for the multi-modal network and enable collaboration among stakeholders have given rise to several research issues. These research issues include acceptability for consumers to pay for the extra cost incurred due to collaboration activities for MaaS systems, modeling of demand, planning for mobility services considering integration with mass transit or other public transport modes, evaluation of MaaS systems via simulation, effectiveness of bundling transportation services, integration of one transport mode in a MaaS system, integration of one transport mode with another one, impact assessment of a transport mode through simulation, integration of the societal goals into MaaS services and methods to support formation of an alliance. Some of these research issues have been studied in the literature in recent years. These research questions along with the studies that respond to them are as follows.

(1)

Question: Are consumers willing to pay for collaboration activities arising from institutionally facilitated interaction between operators?

Results: The study [105] analyzed integrated travel choice experiments based on random parameters logit models. The study showed that consumers would prefer collaboration of operators, engage in integrated transport solutions and be willing to pay for collaboration activities as long as better door-to-door travel options for their whole journey can be delivered.

(2)

Question: How to model demand and plan for ride-shared mobility services running in integration with mass transit?

Results: In the study [80], the multi-agent transport simulation platform, MATSim, was used to model two mobility services. The simulation results for the trip chains aggregated mobile phone network dataset with 722,752 agents generated gain a better insight into users’ travel patterns. The proposed model maximizes the number of users served and quantifies the benefits of integrating with public transport services to enable modal shift from private cars.

(3)

Question: How to evaluate the integrated behavior of MaaS?

Results: In the study [81], a mechanism was proposed to integrate event-based simulator based on attaching/detaching existing simulation models as well as mobility services through application programming interfaces to evaluate MaaS. The mechanism includes independent generation of transportation demands and simulation of user preferences. The effectiveness of the proposed mechanism was demonstrated by applying it to one scenario with private and small area MaaS and the other scenario with public and middle size area MaaS to evaluate the transportation performance. The proposed mechanism enables realizing and evaluating practical MaaS through improving user experiences and quality of services.

(4)

Question: Can consumers benefit from bundling transportation services?

Results: In [115], experiments were performed to estimate and compare consumers’ willingness to pay for stand-alone transportation services with service bundles. One finding of the study was that bundled public transportation, car sharing, and park and ride services were valuated significantly higher than a stand-alone service. Another finding of this study was that bundled bike/e-bike sharing and taxi services was valuated lower. Potential users were highly willing to pay for a smartphone application integrating the services, booking and payment.

(5)

Question: How to integrate freight transport into a Mobility-as-a-Service (MaaS) environment?

Results: Capacity utilization of transport systems in cities might contribute to the improvement of passenger transport and freight movement efficiency and sustainability. To assess the potential impact of potential integration service models, relevant logistics segments were identified to propose service models and evaluate them from the perspective of multi-stakeholder and sustainability in the study [109]. The analysis shows the fitness between on-demand freight and passenger systems combined in MaaS.

(6)

Questions: How to achieve operational, informational and transactional integration required for MaaS to deliver an exceptional user experience to rival the private cars? What factors affect how people make travel choices for adoption of any MaaS offer?

Results: The paper [110] focusses on the user perspective, offers and prospects of MaaS. The paper examines what is understood to date about MaaS and highlights the key point that MaaS is a system beyond the private cars with a mobility intermediary layer. A ‘levels of MaaS integration (LMI) taxonomy’ is presented to capture a hierarchy of user need, including cognitive user efforts, operational concerns, information and transactional integration involved in MaaS.

(7)

Question: How to extend the concept of Mobility as a Service (MaaS) to fully integrate public transportation to make public transport an attractive alternative to private transport?

Results: By extending smart ticketing systems in airline industry with governance and operational processes, fully incorporating operators and utilizing their self-interest to deliver commercially viable and attractive integrated public transport services to consumers, the concept of Collaboration as a Service was proposed in the paper [111] to fully integrate public transportation.

(8)

Question: How to systematically support alliance formation in MaaS?

Results: MaaS relies on integration of various forms of public and private transport services based on collaboration of transport service providers. Forming an alliance is an effective approach to facilitate collaboration of transport service providers. In the study [106], a conceptual model was proposed and ten fundamental propositions for alliance formation were formulated to offer MaaS systems. The proposed alliance formation model was applied in a MaaS pilot in the Netherlands. The proposed alliance formation model and the propositions provide the necessary prerequisites for the design of a governance structure to provide MaaS services.

(9)

Question: How can a public transport authority develop MaaS in rural areas through integrating a public transport service with carpooling?

Results: MaaS involves the integration of different travel modes into an integrated transport service to handle bookings and payments for individual trips. The study [112] presents how a public transport authority developed a MaaS in rural areas through integrating a public transport service with carpooling throughout the development of a project

(10)

Question: How to simulate autonomous mobility-on-demand systems to evaluate its impact from the perspectives of travelers, operators and the city?

Results: In [82], simulation of autonomous mobility-on-demand systems was achieved through co-simulation of two agent-based simulators, MATSim and IMSim, with the former generating transport demand and allocating travelers according to their preferences whereas the latter providing operational execution of transport fleets. The simulation method was applied to probe into the challenges to offer ridesharing services with autonomous vehicles.

(11)

Question: What are the characteristics of a MaaS service and how to compare different services, understand the potential effects of MaaS, and integrate the societal goals into MaaS services?

Results: In the study [113], a topology of MaaS was proposed as a tool to characterize and discuss MaaS. The proposed topology consists of MaaS Levels 0 to 4 to characterize different degrees of integration, where Level 0 denotes no integration; Level 1 denotes integration of information; Level 2 denotes integration of booking and payment; Level 3 denotes integration of the service offer, including contracts and responsibilities; Level 4 denotes integration of societal goals. The levels were used to describe their added value, technical requirements and implications for society, business and users/customers.

(12)

Question: How to assess the effects of ridesourcing services on service providers and the transport system by simulation?

Results: The study [83] presented how to integrate sharing and non-sharing based ridesourcing services into mobiTopp, an agent-based travel demand model by including a vehicle allocation and fleet control component and extending the mode choice by the ridesourcing service to assess the effects of ridesourcing services on service providers and the transport system. In the study, the results concerning performance indices for providers (e.g., the number of bookings, trip times, and occupation rates and influence on other travel modes) for up to 1600 vehicles were analyzed. The analysis is helpful to gain insights into interdependencies between ridesourcing services and other travel modes and paves the way for the planning and regulation of ridesourcing services.

(13)

Question: How to integrate one-to-many peer-to-peer ridesharing with public transit on a MaaS platform for morning commute?

Results: The study [114] focuses on integrating one-to-many peer-to-peer ridesharing and public transit for morning commute. Due to the computational complexity to optimize matching, routing and scheduling for the users in a large scale network, a decision model for the integrated matching problem and an iterative distributed optimization algorithm based on decomposing the original problem into small subproblems within clusters through an incremental approach were proposed. The results of experiments with real-world data show that the proposed algorithm is able to find quality solutions for MaaS systems while significantly reducing the users’ vehicle miles travelled.

(14)

Question: What are the consumers’ intention to subscribe to MaaS, preferences for bundling in MaaS and willingness to pay for extra features of MaaS?

Results: In [116], a method to estimate the latent demand for MaaS based on a stated preference survey and a choice model was proposed. To study the effects of various factors such as service attributes, socio-demographics and transportation characteristics on the decision to subscribe MaaS, a binary mixed logit model was used. Based on the proposed model, the effects of transportation mode pricing schemes, the cross effects between transportation modes and the effects of individual characteristics on the choice of which transport modes to be included in the subscription were studied. The results show that consumers are not yet inclined to subscribe to MaaS. The subscription intention is affected by the service attribute characteristics, the subscription price, and the social influence variables. The most preferred transportation mode is public transportation. Socio-demographic profiles and individuals’ transportation characteristics are critical factors influencing the decision to subscribe and the choice of which transport modes to be included in the bundle.

(15)

Question: How to evaluate added values of MaaS bundles for users with heterogeneous subscription willingness?

Results: In [117], an integrated choice and latent variable model was developed to take into account the factors that influence user subscription to MaaS bundles and estimate the users’ subscription willingness. Four bundles were considered in the study, Bus First, Metro Access, Value Taxi, and Ultra Access. By grouping the users with different subscription willingness to MaaS bundles, the added value of MaaS bundles for different group of users was evaluated. The proposed method was tested based on data collected from a stated preference survey. The findings of this study showed that the target users of Bus First and Metro Access bundles were similar. The target users of Value Taxi and UltraAccess bundles were also similar. In terms of the target users’ subscription willingness, the Bus First and MetroAccess bundles were considerably higher than the other two bundles. The results showed that BusFirst bundle and Metro Access bundle were generally accepted by users.

(16)

Question: How do different MaaS bundling and pricing schemes contribute to sustainable transportation?

Results: The study [118] assessed the influence of applying different bundling and pricing schemes in MaaS on sustainable transportation. A stated portfolio choice experiment was conducted to capture individuals’ intended choice of MaaS. To study the bundles of transportation modes selected by individuals, a mixed logit model was used. The study explores the transportation modes that individuals prefer to include in the bundle, influence of pricing schemes on the bundle as well as the impact on sustainable transport caused by pricing schemes. The findings show that the MaaS’s contribution to improvement of sustainable transport is a non-linear function of decreasing monthly subscription fees and/or increasing length of the subscription.

(17)

Question: How to allocate costs and determine prices offered by operators according to user route choices and operator service choices in MaaS systems?

Results: An algorithm was developed in [119] for a model proposed to allow travelers to make multi-modal trips offered by multiple operators to efficiently generate stability conditions. Application of the model to handle pricing responses of MaaS operators to technological and capacity changes, government acquisition, consolidation, and firm entry was demonstrated through computational experiments. Results showed that the solution time of the proposed algorithm is reduced by 98% on average.

(18)

Question: How to solve the vehicle dispatching service pricing and demand problem in MaaS?

Results: In [120], a two-stage Stackelberg game under different pricing schemes was proposed for the vehicle dispatching service pricing and demand problem. Two different pricing schemes were considered, the independent pricing scheme (IPS) and the competitive pricing scheme (CPS). The vehicle supply–demand relationship and market competition among the modality switch points (MSPs) were also considered in the problem. The paradigm of leaders-followers is used with the MSPs as leaders to set their service pricing strategies first, and the passengers as the followers to determine their service demands. Due to the complexity of the dynamic MaaS market environment, a multi-agent deep reinforcement learning (MADRL) algorithm was developed to achieve the Nash equilibrium (NE) (the optimal pricing and demand strategies) of the formulated game. Convergence to the optimal solution by applying the proposed algorithm was analyzed. The results show that the proposed algorithm outperforms other benchmark schemes under both IPS and CPS in terms of maximizing MSPs’ revenue and passengers’ benefits.

(19)

Question: How to balance sustainability and profitability in electric Mobility-as-a-Service ecosystems by providing carbon emissions reduction incentives?

Results: An ecosystem enabling platform to leverage carbon reduction funds to incentivize travelers to choose electric MaaS services was presented in [121]. A multi-leader multi-follower game model was proposed in [121] to model the interactions between multiple leaders and multiple followers, where leaders represent MaaS platforms and followers represents travelers. A leader competes with other leaders to maximize his/her profits by making bundle allocation/pricing decisions based on the levels of travelers’ participation. A follower minimizes his/her travel costs based on determination of the participation levels for multiple MaaS platforms. An algorithm was developed to solve the proposed multi-leader multi-follower game. The convergence and robustness properties of the proposed algorithm were illustrated by experiments based on real-life data in Australia. For MaaS operators, the proposed method provides valuable insights into balancing profitability with environmental responsibility.

(20)

Questions: Can a MaaS system ensure that travelers and service operators will be better off in terms of welfare while maintain profitability of the platform? Does there exist a pricing scheme that can ensure stability condition under which operators and travelers of a MaaS system are willing to participate?

Results: To answer these questions, Yao and Zhang regarded a MaaS platform for a multi-model transportation network as an intermediary between travelers and service operators in the study [122]. The MaaS intermediary acquires capacity from the service operators and sells capacity to travelers to according to the origin–destination (OD) pairs of their trips. They proposed a model for the assignment problem and the pricing problem for the on-demand mobility services of a MaaS platform in which travelers can plan their trip flexibly. The authors conducted analysis for the many-to-many stable matching issue in a MaaS platform through decomposition of the original problem into an assignment problem and a pricing problem. A penalty-based solution algorithm was proposed for the assignment problem by making use of market clearance condition. A stability condition under which operators and travelers of a MaaS system would be willing to participate was derived. The authors established a stability condition under which operators and travelers of a MaaS system would be willing to participate. The stability condition can be checked without path enumeration.

(21)

Questions: How to develop an insurance product that can cover the policyholders throughout their entire trips with different transport modes? How to incentivize sustainable multi-modal transport through an innovative risk assessment and transfer scheme?

Results: The shift in mobility trends from single modal transport to multi-modal transport creates new demand and opportunity in the insurance sector. The characteristics and operations of multi-modal transport will go beyond any third-party liability cover. Obviously, an all-inclusive insurance product to cover the policyholders throughout their entire multi-modal trips with different transport modes is required. In the study [123], a novel transport insurance product for multi-modal mobility called Pay-As-You-Move was proposed based on usage-based insurance by collecting data from smartphone telematics to detect mobility patterns and assess the risk of an accident. A scheme was proposed to assess the risk of a multi-modal trip by calculating the product of the mileage, the accident probability and the accident severity per mobility type. The proposed novel usage-based insurance product helps promote dependable and greener multi-modal mobility.

7.3. Potential Future Research Directions

The above review of the recent papers related to integration and pricing issues in the context of multi-modal mobility indicates that there are several research issues or questions to be studied. These research issues or questions include (i) methods to orchestrate collaboration for alliance formation of MaaS, (ii) generic simulation methods for MaaS, (iii) systematic approaches to achieve different level of integration and (iv) effective approaches to bundling and pricing transportation services for MaaS. Several potential future research directions based on the review of the papers in the previous subsection are summarized in this subsection.

(i)

Methods to orchestrate collaboration for alliance formation of MaaS

Although the alliance formation model and the propositions proposed in the study [106] have illustrated the necessary preconditions for designing a governance structure to provide MaaS services, it was only applied in a MaaS pilot in the Netherlands. The proposed conceptual model and resulting propositions should be refined and tested in other MaaS-applications to confirm the validity of the model. Different business models need to be examined rigorously based on game theoretical approaches to identify which actors takes, or should take, the lead in alliance formation of MaaS. The study [105] shows that consumers are willing to pay for collaboration activities as long as it can deliver them better door-to-door travel options across the whole journey. Questions such as (1) How to orchestrate such collaboration? (2) How are the institutional rules created to make consumers understand? (3) How are the institutions themselves to be formed? (4) What competing goals and pressures are internal to the different operators and operator types, and how these can be resolved through network governance and management to allow greater participation? and (5) How to bridge between operators with inconsistent goals? need to be studied further. For example, the requirement for profitability creates tension between publicly owned and privately owned operators and this affects their collaborative activities. Additional collaboration mechanisms such as frequent rider incentives and consideration of group travel rather than individually ticketed travel need to be studied.

(ii)

Generic simulation methods for MaaS

Although some of the issues have been addressed in the studies [80,81,82,83] based on simulation approach, there are limitations in these studies. For example, the study [80] lacks technical support if additional components need to be implemented. The methodology was applied to Bristol only. It needs to be validated for large complex scenarios to assess coopetition mechanism for providing seamless integrated door-to-door journeys. For the paper [81], further studies are required to evaluate more complex scenarios, verify the influence of differences in simulators’ resolutions on performance, and introduce shared infrastructures. Although the proposed co-simulation method in [82] enables the assessment of a wide variety of the effects for multiple stakeholders of autonomous mobility-on-demand services such as optimum vehicle fleet size, the vehicles’ occupancy, waiting time, detour time for each passenger, congestion, and the environment, transferability of the co-simulation method to other cities need to be validated.

There are several limitations in the results of the presented scenario in [83], which serves as a technical demonstrator. The results of the study [83] were based on the same decision parameters for the mode choice of public transport and ridesourcing services. To assess the realistic situation, these parameters must be replaced by data collected from surveys covering the usage of ride-hailing and ride-pooling vehicles to obtain more realistic results concerning demand changes and mode shifts.

(iii)

Systematic approaches to achieving different levels of integration

Although the papers [110,113] have characterized different degrees of integration by defining five levels of MaaS integration, how to achieve each level of MaaS integration is not fully addressed in existing studies. For example, the study [109] addressed the issue to integrate freight transport into a Mobility-as-a-Service (MaaS) environment, the paper [111] addressed the issue to fully integrate public transportation, the study [112] presented how a public transport authority developed a MaaS for rural areas by integrating a public transport service with carpooling and the study [114] illustrated how to integrate one-to-many peer-to-peer ridesharing and public transit for morning commute for MaaS. There still lack systematic approaches to achieving different levels of integration for MaaS. Obviously, this calls for more studies on the development of methodologies to effectively achieve each level of integration for MaaS.

(iv)

Effective approaches to bundling and pricing transportation services for MaaS

The results obtained in the study [115] were based on a survey and analysis with a mixed logit model. Although experiments were performed in the study [115] to estimate and compare consumers’ willingness to pay for stand-alone transportation services with service bundles to assess the benefits for consumers, the study was limited. Investigation of the sensitivity of the willingness to pay of consumers with respect to different app features/presentation requires a further study. Conduction of similar experiments in different geographical areas need to be performed to assess the robustness and validity of the results. The study of the effects of flat-rate on bundling transportation services is an interesting issue. The effects of other factors (such as mobility behavior, socio-demographic and household characteristics) on bundling transportation services are interesting research questions to be studied.

Although a method to estimate the latent demand for MaaS based on a stated preference survey and a choice model was proposed in [116], how to combine the proposed model with a simulation technique to capture the dynamics affecting the potential demand for MaaS is an interesting issue. The factor of subsides and the pricing schemes based on the choice of transportation modes to include in the bundle require further studies. Although an integrated choice and latent variable model has been developed in [117] to capture the factors that influence user subscription to MaaS bundles and estimate the users’ subscription willingness, to improve the estimation of the added value of bundle, conduction of surveys with more detailed local population data is required. The study [118] assessed the influence of applying different bundling and pricing schemes in MaaS on the sustainable transportation. An extended study is required to analyze differences between areas, considering factors such as the degree of congestion, ease of finding a parking lot, parking price and public transportation. The factors due to changes in the use of transportation modes and activity-travel patterns and the associated changes in emissions were not taken into consideration in the study [118]. Incorporation of behavioral data is required to elaborate the analysis of the effects of MaaS on sustainable transportation.

An algorithm was developed in [119] for a model proposed to allow travelers to make multi-modal trips offered by multiple operators to efficiently generate stability conditions. However, the issue addressed in the study was based on a static problem formulation. An interesting issue is to extend the study to address day-to-day operations for a dynamic problem setting. Calibration of the model based on real data is also an important issue in applying the proposed method. The proposed assignment game and algorithm has the potential to be applied to other areas such as freight, airlines, other two-sided markets. Applying the proposed algorithms in these areas are interesting but non-trivial research issues.

In [120], a two-stage Stackelberg game under different pricing schemes was proposed for the vehicle dispatching service pricing and demand problem. Analysis of a highly competitive game between mobility service providers and passengers in the MaaS market is one challenging issue. To ensure trust, consideration of blockchain for secure and transparent trading is another interesting issue.

A multi-leader multi-follower game model has been proposed in [121] to model the interactions between multiple leaders and multiple followers, where leaders represent MaaS platforms and followers represents travelers. An algorithm has been developed to solve the proposed multi-leader multi-follower game. However, network configurations were not considered in the model. The influence of network configurations on design of MaaS bundle, allocation of resource, and the efficiency is an important issue. Another research direction is to study the effects of various incentives on traveler behavior, platform competition, and systemic sustainability. Integration of the data-driven MaaS system with the power system or smart grid technologies to optimize the charging infrastructure for different travel modes and enhance management of energy and sustainability is also an interesting research direction.

Although [122] have established a stability condition under which operators and travelers of a MaaS system are willing to participate, the condition is based on several assumption. There are several future research directions related to [122]:

(1)

Incorporating the operators’ decisions on service capacities and MaaS fares into the model.

(2)

Relaxing the fixed total demand and analyzing the robustness of a MaaS system with respect to the future growth of travel demand.

(3)

Considering a realistic setting with sophisticated operational strategies and heterogeneous access time, instead of uniform distribution of vehicles and bikes.

(4)

Exploring the impact of topological properties of the multi-modal transportation network on the social cost saving and profitability of MaaS.

(5)

Considering multi-objective optimization to ensure all stakeholders can be better off.

(6)

Considering wholesale capacity constraint from the perspective of operators and investigating the influence of wholesale capacity constraint on properties and solution algorithms of the model.

(7)

Studying the differences between matching stability and supply-chain equilibrium and investigating whether matching stability and supply-chain equilibrium fits better with real practice.

The idea of “Pay-As-You-Move” insurance product for multi-modal mobility in the study [123] sparks two future research directions: (1) considering contextual information related to policyholders’ trips, e.g., time of day, weather, and light conditions, in assessing the risk of different mobility types to enhance the underwriting information; (2) developing the pricing structures for the novel insurance by modifying traditional insurance pricing structures to meet the need of a multi-modal insurance policy.

8. Discussion

This review is centered on the papers published in the Web of Science that are related to “mobility” and “multi-modal” in recent years. To systematically review and categorize existing works related to multi-modal mobility, the keywords “mobility” and “multi-modal” were used in the initial search. The results showed that the number of published papers grew from 34 (during the years 2006 to 2015) to 185 (during the years 2016 to 2024). The exponential growth in the number of published papers related to “mobility” and “multi-modal” over the years shows that many emerging research directions have attracted the attention of related research communities. The results of the initial search were analyzed using VOSviewer to identify terms or keywords for refining the search. We classified papers into five categories by combining different terms or keywords to perform the refined search. The first category of works was obtained by searching with string AB = (“MaaS” AND “multi-modal”) and the second category was obtained by searching with string AB = (“sustainable” AND “mobility” AND “multi-modal”) or string AB = (“sustainability” AND “mobility” AND “multi-modal”). The third category of works was obtained by searching with string AB = (“shared mobility” AND “multi-modal”) or string AB = (“ridesharing” AND “multi-modal”). The fourth category of works was obtained by searching with strings AB = (“bike” AND “transport” AND “multi-modal”), AB = (“micromobility” AND “multi-modal”) and AB = (“scooter” AND “transport” AND “multi-modal”). The fifth category of works was obtained by searching with the query strings AB = (“MaaS” AND “integration”) and AB = (“MaaS” AND “pricing”). Content analysis was performed after the refined search. If a paper was not appropriate for a category, it would be reassigned to another category that is more relevant. This review was based on the five categories obtained. The number of papers is 17 in Category 1, 19 in Category 2, 10 in Category 3, 11 in Category 4, and 21 in Category 5. Category 5 ranks highest and Category 2 ranks second highest in terms of the numbers of papers.

In the remainder of this section, we will first analyze research topics and issues identified in this review and then discuss opportunities and challenges for future research.

8.1. Analysis of Research Topics and Research Issues

To review the papers efficiently and systematically, we identified the research topics and research issues. Figure 9 shows the research topics. There are connections between different topics. For example, studies of adoption factors might require modeling and/or simulation on one hand. Therefore, there is a connection between the topic “adoption factors” and the topic “modeling, simulation and planning”. One the other hand, studies of modeling, simulation and planning typically relies on the knowledge of adoption factors to develop the simulation models or planning software. Another example, studies of integration might require modeling and/or simulation. Knowledge of integration is required to construct modeling, simulation and planning tools. Figure 9 shows the connections between the topics of the literature in Table 3.

Table 9. Research issues related to the topic “modeling, simulation and planning”.

Modeling, Simulation and Planning
Advanced frameworks/models to simulate/optimize the performance of MaaS	[65,66,67,68]
Modeling, simulation and planning in multi-modal transport networks	[69,70,71,72,73,74,75,76,77,78,79]
Generic simulation methods for MaaS	[80,81,82,83]

,

Table 10. Research issues related to the topic “adoption factors”.

Adoption Factors
Enabling factors/barriers of modal shift to sustainable transport	[46,47,48,49,50,51]
Factors influencing adoption of sustainable mobility behaviors	[52,53,55,56]
Approaches to promoting sustainable mobility	[54,57,58]
Characteristics and adoption factors of micromobility in multi-modal transport	[59,60,61,62,63,64]

,

Table 11. Research issues related to the topic “integration”.

Integration
Systematic approaches to achieving different levels of integration	[109,110,111,112,113,114]
Effective approaches to bundling and pricing transportation services for MaaS	[115,116,117,118,119,120,121,122,123]

,

Table 12. Research issues related to the topic “impact”.

Impact
Analysis of the potential impact caused by MaaS	[84,85,86]
Factors with impact on transport modes and sustainability	[87,88,89]
Studying the relation between different transport modes and the impacts of transport	[90,91]
Impacts and substitution effects of micromobility	[92,93,94,95,96]

,

Table 13. Research issues related to the topic “implementation”.

Implementation
Strategies to facilitate the implementation of multi-modal shared mobility	[97,98,99]
Effective approaches to dealing with threats while protecting privacy and allowing data	[100]
Urban policy design and implementation methods for sustainable urban mobility	[101,102]
Leveraging the benefits of shared mobility to enhance multi-modal transport	[103,104]
Methods to orchestrate collaboration for alliance formation of MaaS	[105,106]

and

Table 14. Research issues related to the topic “performance”.

Performance
Performance evaluation in multi-modal transport systems	[107,108]

show the related research issues for each topic. In each of these figures, the research issues are listed in the left column while the related papers are listed in the right column.

Through content analysis, several research issues and topics were identified in each category. These research issues and topics are summarized in

Table 15. Number of papers for each research issue.

Category	Issue	No. of Papers	Topic
1	(i) Enabling factors/barriers of modal shift to sustainable transport	6	Adoption factors
	(ii) Analysis of the potential impact caused by MaaS	3	Impact
	(iii) Strategies to facilitate the implementation of multi-modal shared mobility	3	Implementation
	(iv) Advanced frameworks/models to simulate/optimize the performance of MaaS	4	Modeling, simulation and planning
	(v) Effective approaches to dealing with threats while protecting privacy and allowing data sharing in MaaS	1	Implementation
2	(i) Factors influencing adoption of sustainable mobility behaviors	4	Adoption factors
	(ii) Approaches to promoting sustainable mobility	3	Adoption factors
	(iii) Factors with impact on transport modes and sustainability	3	Impact
	(iv) Modeling, simulation and planning in multi-modal transport networks	5	Modeling, simulation and planning
	(v) Urban policy design and implementation methods for sustainable urban mobility	2	Implementation
	(vi) Performance evaluation in multi-modal transport systems	2	Performance
3	(i) Modeling, simulation and planning in multi-modal transport networks	6	Modeling, simulation and planning
	(ii) Leveraging the benefits of shared mobility to enhance multi-modal transport	2	Implementation
	(iii) Studying the relation between different transport modes and the impacts of transport modes	2	Impact
4	(i) Impacts and substitution effects of micromobility	5	Impact
	(ii) Characteristics and adoption factors of micromobility in multi-modal transport networks	6	Adoption factors
5	(i) Methods to orchestrate collaboration for alliance formation of MaaS	2	Implementation
	(ii) Generic simulation methods for MaaS	4	Modeling, simulation and planning
	(iii) Systematic approaches to achieving different levels of integration	6	Integration
	(iv) Effective approaches to bundling and pricing transportation services for MaaS	9	Integration

. Note that there are 19 papers related to “modeling, simulation and planning in multi-modal transport networks” in research issue (iv) of Category 1, research issue (iv) of Category 2, research issue (i) of Category 3 and research issue (ii) of Category 5. Therefore, the research topic “modeling, simulation and planning in multi-modal transport networks” rank highest in terms of the numbers of papers. Note that there are also 19 papers related to “adoption factors in research issue (i) of Category 1, research issue (i) and research issue (ii) of Category 2 and research issue (ii) of Category 4. Therefore, the research topic “adoption factors” also rank highest in terms of the numbers of papers. There are 15 papers related to “integration” in research issue (iii) and issue (iv) of Category 5. Therefore, the research topic “integration” also rank third in terms of the numbers of papers. The number of papers in research issue (iv) “Effective approaches to bundling and pricing transportation services for MaaS in Category 5 is 9 and also ranks high. Therefore, the research issue “Effective approaches to bundling and pricing transportation services for MaaS” is also an emerging research direction.

The rightmost column of Table 15 shows the research topics. The papers can be grouped according to the topics. The number of papers for each topic is shown in

Table 16. Number of papers by research areas.

Topic	No. of Publications
Modeling, simulation and planning	19
Adoption factors	19
Integration	15
Impact	13
Implementation	10
Performance	2

. The results in Table 16 indicate that the topic with the highest number of publications in the past 8 years is “modeling, simulation and planning”. Another topic with the highest number of publications is “adoption factors”. The topics “integration” and “impact” are ranked third and fourth, respectively. The topic “implementation” is ranked fifth. The lowest ranking topic is “performance”. The above statistics only show the trends of the past 8 years during which MaaS was still in its infancy. With more pilots and MaaS systems being implemented, transport planners, service providers, researchers and practitioners will be able to develop best practices, principles, strategies, optimization tools and incentive programs to achieve sustainable, accessible, seamless and efficient door-to-door multi-modal transport to meet SDGs challenges based on the previous implementation experiences of MaaS.

8.2. Opportunities and Challenges

Several opportunities, challenges, and research gaps identified in this review are discussed below. In Table 15, the total number of papers related to the research topic “modelling, simulation and planning” is greatest. Transport system models play a pivotal role in the development of MaaS with multi-modal transport as they provide the capabilities to study and optimize MaaS in real urban areas. In the literature, several modelling frameworks for the design of sustainable transport services have been proposed. For example, a modelling framework called Sustainable MaaS for achieving sustainability objectives and goals has been proposed in [134]. A hypothesis on the evolution of MaaS divides the evolution of MaaS into several stages, N-MaaS (No MaaS), MaaS 1.0, MaaS 2.0 and MaaS 3.0 [134]. N-MaaS refers to separate services without integration. MaaS 1.0 refers to integrated transport services with ICT based platform for MaaS and user acceptance. MaaS 2.0 includes the features of MaaS 1.0 and transport system models to design and manage the system with the DSS platform. MaaS 3.0 includes the features of MaaS 2.0 and takes into account the sustainability objectives, targets and goals with Space Economic Transport Interaction (SETI) models [135] and Environmental Impact Functions (EIFs) [136], where the SETI model consists of a Spatial economic macro-model and a transport macro-model to capture the interaction between spatial economic and transport. The framework proposed in [134] includes the general optimization formulation of the design models detailed in [137], a decision support system (DSS) platform and sustainable objectives and goals. The work of [137] reports a methodology that consider actors such as operators, companies, users, citizens and system manager/planner in MaaS to specify transport system models for estimating the effects of decision makers’ actions on MaaS. The methodology includes a framework to support S-MaaS policies definition and transport system models for services design. The optimization model specified in [137] includes the objective function, control variables, internal variables, external/technical constraints and behavioral constraints. The DSS platform refers to models, algorithms and methods that support stakeholders such as public authorities, operators, companies and users.

In addition to the optimization frameworks mentioned above, modeling of travel demand and its interaction with supply is an important part in the development of a complete methodology. The study reported in [136] focuses on travel demand modelling and analysis based on the transportation system models. Different means for collecting and estimating the travel demand may be estimated by means of disaggregated sample, surveys and aggregate surveys such as characteristics of the area and traffic flows. The dynamic model to capture the interaction between demand and supply in MaaS with ICT has been studied in [138] to achieve goal of sustainable mobility. In [138], the authors analyzed the goals and targets of Agenda 2030 and highlighted those for which MaaS can contribute to the pursuit. They formulated a dynamic model based on transport system models which includes the formalization of emerging ICT technologies in the demand model, the supply model and the interaction model. The authors applied the proposed transport system models to quantitatively evaluate the MaaS development policies with respect to the targets of Agenda 2030. In addition to the above studies, a project that focused on the development of methodologies for the evaluation of MaaS, conduction of pilot surveys on samples of potential users, and analysis of results to assess the potential of MaaS and its impacts on mobility was reported in [139].

Although the studies mentioned above provide systematic methods to pave the way for the development of sustainable MaaS, a research gap remains in applying the proposed methods to solve real-world multi-modal mobility optimization problems, as this is a non-trivial task. For example, to apply the optimization model specified in [137], the objective function, control variables, internal variables, external/technical constraints and behavioral constraints that need to be defined according to the target application scenario. The complexity of single modal shared mobility problems are already very complex [140,141]. The complexity to develop solution algorithms for the optimization sustainable multi-modal MaaS is one challenge. Other issues such as scalability and agility of solution algorithms pose additional challenges to develop solvers for the optimization and planning of sustainable multi-modal MaaS.

In Table 15, the number of papers addressing the research issue “Effective approaches to bundling and pricing transportation services for MaaS” is the second highest among all research issues. This shows that bundling [118] and pricing [119,122] transportation services have received prominent attention. The number of papers addressing research the issue “Systematic approaches to achieving different levels of integration” is also high. An interesting future research direction to close the research gap related to this issue is to study how to achieve different levels of integration through developing proper bundling and pricing schemes to incentivize the potential MaaS users and service providers. Pricing in emerging single modal mobility services has been extensively studied. A comprehensive review of pricing in three emerging systems: ridesharing, carsharing, and shared autonomous vehicles (SAVs) can be found in the paper [142]. In [142], the authors categorize the pricing approaches in single modal mobility services into trip-based pricing, origin-based pricing, duration/distance-based pricing, and auction. However, an in-depth study is still lacking on how to apply or extend these single modal mobility service pricing approaches to address pricing issues in MaaS.

A research issue closely related to pricing is cost allocation or allocation of benefits. The issue of cost allocation or allocation of benefits has been addressed in the context of single modal transport, but it has been less studied in the context of multi-modal transport. Cost-sharing mechanism design for ridesharing has been studied in [143]. The cost-sharing problem in the context of ridesharing refers to the problem to determine the allocation of the total ride cost between the driver and the passengers. The setting of the problem considered in [143] assumes that each driver has his/her own origin and destination. Several desirable properties of cost-sharing mechanisms have been identified and a framework to create specific cost-sharing mechanisms in the context of ridesharing has been proposed. However, whether the cost-sharing framework proposed in [143] can be applied to multi-modal transport systems requires further studies. In a single modal shared mobility system, how to allocate benefits to the drivers, passengers and service providers will influence the satisfaction of participants. Several schemes have been proposed to address this issue in [144,145,146]. In [145,146], three schemes have been analyzed to compare their influence on satisfaction of participants. Another interesting issue that has been studied in single modal shared mobility system is to attract participants through the incentive of discount guarantee. For example, in [147,148], different metaheuristic algorithms have been proposed to solve the ridesharing problem with discount guarantee. However, there lacks a study on how to provide an incentive with discount guarantee in MaaS systems. Cost allocation for ridesharing services considering fairness issue has been studied in [149]. However, whether the above studies can be extended to deal with MaaS requires further investigation to bridge the gaps identified above.

The three research issues—“enabling factors/barriers of modal shift to sustainable transport”, “characteristics and adoption factors of micromobility in multi-modal transport networks”, and “factors influencing adoption of sustainable mobility behaviors”—account for a total of 16 papers. This indicates that adoption factors and barriers in the context of multi-modal transport were extensively studied in the past 8 years. The reason adoption factors and barriers were extensively studied is that they are closely related to critical success factors and critical failure factors, respectively. The former refers to essential elements that must be executed well to achieve strategic objectives, while the latter refers to elements that might prevent a project or organization from reaching its goals. Although the number of papers related to the research issue “Approaches to promoting sustainable mobility” is small, the development of innovative schemes to attract potential MaaS users and enhance their participation through rewards and incentive programs remains an important topic for future research and a gap to be filled.

When a new paradigm introduces a service that relies on existing ones, those existing services and their stakeholders may be affected by changes in market size and by the need to adapt how the old services operate in order to accommodate the new one. The three research issues—“impacts and substitution effects of micromobility” [92,93,94], “analysis of the potential impact caused by MaaS” [84,85,86], and “studying the relation between different transport modes and the impacts of transport modes” [90,91]—focus on the studies of the impacts caused by the emerging transport modes, their interaction and substitution effects. Understanding the impacts will be helpful for transportation planners to develop a smooth and feasible path in the migration and evolution processes to achieve sustainable multi-modal mobility. Analysis of the impacts caused by the emerging transport modes and their interaction with existing transport modes relies on the development of analysis or simulation methods/tools that can capture the dynamics and characteristics of real multi-modal MaaS systems. However, there are limitations in the existing analysis or simulation methods. For example, the spatial agent-based model used in [92] does not consider weather factors or psychological aspects such as comfort, perceptions of safety, and leisure travel. The analysis provided in [86] does not take into account transit frequency, vehicle capacity, or hidden costs associated with comfort during trips. These research gaps, due to the limitations of existing analysis and simulation methods, need to be closed.

Simulation is one popular approach to studying MaaS and assessing their impacts on peoples’ travel behavior, modal shifts in cities and bundling/pricing transportation services for MaaS. Different models have been developed to predict modal shifts to sustainable transportation options. These include agent-based models [150], statistical models [95], discrete choice models [60,105,116,118] and system dynamic models. Statistical models and choice models cannot capture the dynamics behind users’ choices and preferences, which are non-linear and evolving features that characterize mobility and behavioral change, and the interactions between agents and environments. System dynamics models is a top-down approach that model the aggregate system instead of disaggregated parts of a system. Agent-based models overcome the shortcomings of statistical models, discrete choice models and system dynamic models as they are a bottom–up approach capable of capture heterogeneous mobility decisions made by individual agents and their interactions. As the agent-based approach allows for the modelling of flexibility, proactivity and negotiation capabilities of entities in collaborative transportation systems such as MaaS, agent-based modelling (ABM) is the mainstream technique for modeling and analyzing the mobility transition [150]. The study [151] indicates that MATSim, an open-source framework for large-scale agent-based transport simulations, is one popular tool used to estimate the transport demand. MATSim supports the co-evolutionary algorithm, which performs optimization within an agent’s set of plan based on the scoring functions of individual agents and is different from evolutionary algorithms that aims to find the global optimal solution. MATSim can be extended with other modeling framework such as BEAM [152], a modeling framework for behavior, energy, autonomy and mobility, to specify multiple mode choices. The MaaS in cities with multi-transport modes can be modeled as “System of Systems” design problems [124]. Such a “System of Systems” relies on the use of two or more simulators to model and simulate the behaviors in multi-modal MaaS. In [82], the concept of co-simulation was proposed to study the interaction of the demand model provided by the MATSim simulator and a fleet simulation model provided by the IMSim simulator. Despite the studies mentioned above, research gaps remain in studying systematic approaches to combining optimization methods with individual simulators and in building an overall model for multi-modal MaaS through the co-simulation of multiple optimization-based simulators.

Although the research issue “performance evaluation in multi-modal transport systems” was less explored, this was due to the lack of theory and advanced tools to support performance evaluation for real multi-modal transport systems. With more MaaS pilot projects being executed, advances in theory and the development of advanced tools for modeling, analyzing, and simulating multi-modal transport systems, the research issue ‘performance evaluation in multi-modal transport systems’ is likely to be addressed more comprehensively in the future.

According to the review in this paper, most studies were conducted for a city or a specific area covering a few cities. Whether the approaches are applicable and the results are transferable to other cities or areas needs to be verified. Effectiveness of an approach to address a specific issue requires a comparative study of applying the approach to different cities or areas. However, a comparative study of applying the same approach to address a specific issue in different cities or areas was less explored in [52,70,82,84,88,107]. For example, in [84], the authors studied how MaaS could impact GHG emissions in an urban setting based on activity-based model for Amsterdam, the Netherlands. Another example, paper [70] addressed the issue about how to develop a framework for modelling various facets of AMoD to demonstrate the benefits of the proposed framework based on a case study in Singapore. The paper [52] explored mobility motives of travelers to find the determinants for travelers to adopt sustainable mobility behaviors based on a sample of the Swiss population. The study [107] probed into the problem about how to estimate long-term travel delay in multi-modal systems with different transport modes based on the Shenzhen city, China. The paper [88] aimed to study the impact of strategies on sustainable mobility for a metropolitan region in the Netherlands including Rotterdam and The Hague by employing activity based travel demand models. The proposed co-simulation method in [82] was applied to assess a wide variety of the effects for multiple stakeholders of autonomous mobility-on-demand services for the Royal Borough of Greenwich (London, UK). The gaps regarding the applicability of existing approaches and the transferability of existing results in [52,70,82,84,88,107] to other cities or areas need to be filled in the future.

Apart from the research issues and research gaps mentioned above, the implementation of multi-modal shared mobility calls for a collaborative partnership between service providers and the government to jointly establish a merit-based business model, improve infrastructure provision, and create smart applications. The requirements of designing environmentally sustainable, economically viable, and socially acceptable urban mobility systems pose challenges in the decision-making processes of cities. The issue is to develop an innovative and collaborative decision framework to support integrated sustainable urban policy design for sustainable urban mobility. Many papers included in this review suggest government intervention or policy implementation to support the development of multi-modal MaaS services. For example, the study [49] indicates that without proper government intervention, the efficiency and equity promises of MaaS are not achievable. The study [98] suggests that government policy makers should innovatively promote shared MaaS through policy implementation. The study [64] provides insight into how governance stakeholders shape sustainability transitions and technology uptake by exploring their perspectives in depth. The study [53] explores the connection between life course and sustainability mobility behaviors of millennials. The study suggests policy interventions to support a sustainable lifestyle. The study [60] recommends the formulation of policies to increase bicycle ownership and improve cycling infrastructure. In [101], a multi-level decision framework was proposed for the design of urban mobility policies based on multi-criteria resource allocation decisions at different levels. Additionally, an interactive scenario-simulation tool was developed in [101] to support decision makers. In summary, governance stakeholders play a pivotal role in taking action toward sustainable development by promoting enterprise and industrial transformation. However, existing studies on the role of governance stakeholders in shaping MaaS uptake are limited. Understanding how governance stakeholders might shape sustainability transitions and technology uptake is an important issue.

The rapidly expanding research issues related to sustainable multi-modal MaaS are both multidisciplinary and interdisciplinary in nature, and they have significant impacts on stakeholders. Achieving the vision of sustainable multi-modal MaaS requires practitioners, experts and stakeholders from different disciplines to draw on their disciplinary knowledge, work collaboratively, and integrate knowledge and methods using a true synthesis of approaches. The collaborative actions of a wide range of stakeholders and practitioners can lead to a better understanding of the concept of MaaS and generate the depth of knowledge [153] needed for its inception, planning, and implementation, based on previous MaaS pilot experiences. For example, several previous MaaS pilots have shown that public-private collaboration or partnership is a crucial element in the realization of MaaS, though it introduces new issues for transport governance [154,155]. Balancing the control of public transport service providers with the interests of private sector mobility service providers poses significant challenges in the development of both urban and rural MaaS.

Previous experiences from MaaS initiatives in cities around the world provide valuable references for the development and implementation of sustainable MaaS. In what follows, we briefly review the MaaS experiences from Helsinki, Singapore, and Shanghai.

Helsinki is a pioneering city in the successful implementation of a MaaS solution [156]. In 2017, the Whim app was launched by MaaS Global. The city’s approach has been studied extensively, offering insights into the factors contributing to its success. Key factors in Helsinki’s MaaS success include: (1) a user-centered app to integrate various transportation modes—public transit, bike sharing, e-scooters, taxis, and car rentals; (2) legislative support requiring transportation carriers to release travel data for planning purposes as well as make ticketing available to third-parties for application development; (3) reliable public transport infrastructure; and (4) significant reductions in environmental impact by shifting users from private cars to public transport and shared services. The early success of the Whim app led to its operation in several countries, including Finland (Helsinki and Turku), Austria (Vienna), Belgium (Antwerp), the United Kingdom (Birmingham), Japan (Tokyo), and Switzerland (various cities nationwide). However, MaaS Global, the Finnish company behind Whim, filed for bankruptcy in March 2024. In April 2024, Dutch mobility platform umob acquired MaaS Global.

There are many studies related to the development of MaaS in Singapore [157,158,159]. The Zipster app, launched in 2019, was the first all-in-one transport app in Asia to integrate various mobility services, including public transit, ride-hailing, bike-sharing, and carsharing [160]. However, it was discontinued after two years due to limited user adoption and profitability concerns. Experts suggest that Singapore’s already efficient and affordable public transport system may have reduced the perceived need for such an integrated platform. Despite the setbacks of the Zipster app, a survey in 2024 indicated that 56% of Singaporeans utilized shared mobility services—such as ride-hailing, carpooling, and bike-sharing—typically two to five times per week. Car ride-hailing services were the most popular, used by 84% of respondents [161].

China’s MaaS initiatives are rapidly evolving, with ongoing efforts to expand and optimize services, build mutually beneficial ecosystems, and reduce carbon emissions across all services. The development of MaaS in China is focused primarily on integrating existing state-run public transport services and secondarily on incorporating additional mobility services beyond conventional public transport. There are several studies related to the research issues of the impact of bike sharing on public transport and the integration with public transport [162,163]. As of 2023, 41 Chinese cities in the developed eastern regions with high political influence and larger populations have hosted MaaS initiatives and associated platforms [155]. For example, Shanghai’s official all-in-one mobility platform called Suishenxing app was developed by the state-owned Shanghai Mobility Service Technology Co. Ltd. (Shanghai, China) and launched in October 2022. This app marked a significant step in the Shanghai’s MaaS initiative as it integrated various transportation services such as public transportation, taxi and vehicle hailing, bike sharing, and smart parking into a single platform. On its second anniversary in October 2024, the Suishenxing app served over 82 million daily trips and provided real-time information across various transportation modes. Suishenxing is primarily a government-led initiative aimed at enhancing urban mobility and promoting green transportation, rather than generating profit. As of now, there is no publicly available data detailing the profitability of Shanghai’s Suishenxing. The current financial performance and profitability of Shanghai’s Suishenxing remain undisclosed.

The stories of the MaaS initiatives in Helsinki, Singapore, and Shanghai can be described by a saying: “Love without bread is bound to wither; whereas bread without love is far too lonesome.” The grand vision of achieving environmentally sustainable, economically viable, and socially acceptable urban mobility represents our “love” for the Earth and future generations. Financial sustainability is the “bread”—a prerequisite for balancing the acquisition of necessary funding with the pursuit of long-term social and environmental goals. In addition to the important issues discussed above, financial sustainability is an essential element for the success of MaaS. Without a sustainable business model, maintaining long-term operations of MaaS can be challenging. Good news is that the study [164] demonstrated that, under certain pricing and assignment strategies, a MaaS platform could achieve profitability while also enhancing traveler welfare and increasing revenues for transit operators. This result indicates that a sustainable business model for maintaining long-term MaaS operations exists. How to systematically develop sustainable business models for MaaS to meet the needs of various cities or countries remains an interesting issue.

9. Conclusions

The pressure to reduce emissions from road transport is creating several emerging transport modes such as shared mobility and micromobility in cities around the world. Hence, the transport systems in many cities are multi-modal and consist of these emerging transport modes as well as the existing traditional transport modes. Coexistence of the emerging transport modes and the existing traditional transport modes presents many research issues and poses challenges in planning, design, implementation and operations of the multi-modal networks in cities. The key issues are to make different transport modes work together effectively and flexibly to provide seamless, accessible, sustainable and efficient mobility services to users. The challenges are due to the complex interactions between different transport modes, the unforeseen impacts on the existing transport modes caused by the emerging transport modes, the lack of effective methods to design/integrate the emerging transport modes in the existing transport modes, the complexity to determine the travel routes considering mode choices and preferences of users, the difficulty in evaluating the performance of multi-modal networks considering travelers’ behaviors and travel types, the deficiency of urban policy design/implementation methods for supporting multi-modal mobility systems and the lack of well-accepted approaches to enabling/encouraging users to adopt sustainable mobility behaviors through providing incentives to promote sustainable mobility.

In this paper, many research questions related to multi-modal mobility in cities studied in recent years have been reviewed. This review shows that the recent studies have made progress to respond to some of the above challenges and paved the way for the further development of theory and practices for planning, designing, implementing and promoting multi-modal mobility systems. However, more efforts must be made to close the gap between the solution methods needed and the ones available in the literature to provide seamless, accessible, sustainable and efficient mobile services to users.

In terms of the number of publications in the past 8 years, the topics “modeling, simulation and planning” and “adoption factors” are ranked second. The topics “integration” and “impact” are ranked third and fourth, respectively. The topic “implementation” is ranked fifth. The lowest ranking topic is “performance”. Although the topic “performance” is ranked lowest, this is due to the lack of sufficient number of multi-modal MaaS systems implemented. We believe that the ranking of the topic ‘performance’ will be improved as more multi-modal MaaS systems are implemented. Based on the above analysis, the findings of this study suggest that more attention should be paid to several directions: (i) effective approaches to modeling, planning and simulation of the emerging transport modes, (ii) optimization/heuristic/metaheuristic approaches to the determination of the travel routes based on mode choices, requirements and preferences of users, (iii) studies of adoption factors based on advanced simulation tools/methods to assess the performance of multi-modal mobility networks under dynamic travel demand, travelers’ behaviors and travel types, (iv) methods to design and integrate emerging transport modes in the existing transport networks, (v) impact and substitution effects of the emerging transport modes, and (vi) design of policy, implementation procedures and effective incentive schemes to attract the stakeholders and facilitate the development and adoption of sustainable multi-modal mobility systems. The review in this study shows that the research issues/questions related to multi-modal mobility are highly diversified. These diverse research issues/questions typically rely on fusion/cross-fertilization of multi-disciplines and pose challenges in the development of viable and effective solution methods to fill the gaps in the multi-modal mobility literature.

Issac Newton said, “If I have seen further it is by standing on the shoulders of giants”. The phrase “standing on the shoulders of giants” was originally used by the 12th-century philosopher, Bernard de Chartres, to express the idea of intellectual progress through building upon earlier works. A lot of technological advancements are based on incremental improvements to existing technologies and the adoption of emerging paradigms to enhance efficiency, productivity, sustainability and overall human well-being. The papers included in this review might be the “shoulders of giants” for researchers to stand and develop new effective methodologies for the realization of sustainable multi-modal MaaS system without reinventing the wheel. However, there are still many challenges to face when one standing on the shoulders of giants. The research issues highlighted in this paper reflect some of these challenges.