Next Article in Journal
Preventive Measures against Ebola and COVID-19 and Their Impact on Human Dirty-Hand Disease Mitigation in Côte d’Ivoire
Next Article in Special Issue
Identification of Solutions for Vulnerable Road Users Safety in Urban Transport Systems: Grounded Theory Research
Previous Article in Journal
Complex Analysis of an Auxetic Structure under Compressive Loads
Previous Article in Special Issue
Backward-Facing Analysis for the Preliminary Estimation of the Vehicle Fuel Consumption
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Research on Parking Recommendation Methods Considering Travelers’ Decision Behaviors and Psychological Characteristics

1
Beijing Key Lab of Traffic Engineering, Beijing University of Technology, Beijing 100124, China
2
School of Transportation, Southeast University, Nanjing 214135, China
*
Author to whom correspondence should be addressed.
Sustainability 2023, 15(8), 6808; https://doi.org/10.3390/su15086808
Submission received: 19 February 2023 / Revised: 11 April 2023 / Accepted: 13 April 2023 / Published: 18 April 2023

Abstract

:
Intelligent parking services can provide parking recommendations and reservations for travelers. They are an effective method for solving the cruising for parking problems in big cities. This research conducted a sequential parking decision behavior survey and analyzed travelers’ parking choices and reservation behaviors at different stages of the travel process. Then, a parking recommendation model was established to consider the travelers’ psychological thresholds and the attention to parking factors. The effects of different parking recommendation schemes in different situations were further explored based on parking simulations. It was concluded that travelers were more willing to accept and use the parking recommendation system. A total of 56% of travelers chose to make parking reservations during the travel process. The satisfaction proportion of the psychological threshold for the parking reservation group was higher than that for the non-parking reservation group. A dynamic parking recommendation scheme with a regulation threshold can change the recommendation strategy according to the overall utilization of parking lots. The implementation of the scheme can not only improve travelers’ parking experience, but it can also effectively balance the utilization of parking resources. It can be applied to different parking utilization situations and produce good performance. The research results provide references for the design and application of intelligent parking services in order to solve parking problems.

1. Introduction

With the improvement of residents’ consumption levels, the rapid growth of motor vehicles brings numerous parking problems. In urban central business districts, there is a shortage of parking supplies, which makes it more difficult for vehicles arriving later to find a suitable parking space at their destination, even if there is no parking space. It makes them have to spend more time searching for parking spaces in other areas, which leads to the phenomenon of cruising for parking. Generally, it takes about 7 to 8 min for people to find a suitable parking space, which reduces their parking efficiency [1]. When the number of cruise vehicles increases to a certain extent, it will have an impact on the surrounding road traffic, leading to traffic congestion, causing vehicles to drive at low speeds, generating more emissions, and increasing environmental pollution [2]. The relevant research shows that 30% of vehicles in the traffic flow cruise for parking [1]. Additionally, due to cruising for parking, the road traffic flow has increased by 25–40% [3]. The annual economic loss caused by cruising for parking in Schwabing, Germany, is estimated at EUR 20 million [4]. Intelligent parking services (IPS) can provide travelers with real-time parking occupancy information by integrating wireless communication technology, mobile terminals, and GPS. Parking recommendations and reservations can help travelers find a free parking space faster and easier, thereby reducing their cruising time, improving parking efficiency, and maximizing parking resources, thus greatly reducing the impact of cruising vehicles on road traffic [5,6]. Therefore, IPS is an effective method to solve parking problems in big cities.
Certain parking reservation systems, such as SpotAngel, can obtain the latest information, which includes the location and pricing, as well as details concerning what free and paid parking spaces are open. It can help travelers find available parking spaces and parking prices, as well as inform them on the opening hours for parking lots. It can also filter the available parking spaces by parking price, payment type, etc. Existing intelligent parking services mostly show the location, number of vacant parking spaces, and parking prices of parking lots within a certain range of their trip destination to travelers. However, little research has been conducted on parking choice behavior, specifically considering the psychological attention and thresholds of individual parking factors. Additionally, it is also rare to consider recommending parking lots from both the perspectives of travelers and managers. Based on the survey and analysis of sequential parking decision behaviors, this research constructs a parking recommendation model and a personal parking decision model, both of which consider the parking factors regarding the psychological thresholds and attention of travelers. Different parking recommendation schemes were designed from the perspectives of travelers and managers, and their applications were analyzed from the perspectives of the parking utilization rate and some other simulation result indicators, ultimately obtaining the optimal parking recommendation scheme. The effects of the optimal recommendation scheme under different initial parking occupancy statuses, parking reservation proportions, and parking regulation thresholds were also explored. The research conclusions can provide a reference for the design and application of intelligent parking systems and help to further reduce a series of urban parking problems.

2. Literature Review

Many scholars have carried out studies on parking choice behavior, as well as parking recommendations and reservations. The studies concerning parking choice behavior mainly focus on the influence of personal and travel-related factors [7]. Zhang and Zhu [8] analyzed the on-street parking choice behaviors in the city center. They found that travelers were willing to pay more when choosing an on-street parking space that is closer to the destination. Khaliq et al. [9] used a mixed multinomial logit model to explore on-street parking choice behaviors and concluded that factors such as parking price, intended parking duration, and parking convenience all had a strong influence on travelers’ on-street parking choices. Soto et al. [10] established a mixed discrete choice model and found that the parking price and walking time after parking were two important factors that affect travelers’ choices for paid and non-paid on-street parking spaces and paid public parking lots. Tian et al. [11] proposed a dynamic parking pricing model to balance the utilization of parking resources. It was concluded that the parking price and walking distance after parking were important influencing factors for parking choice [11,12].
Intelligent parking services can provide parking information near the trip destination for travelers and guide them to park their cars [13]. Cao and Menendez [14] concluded that the adoption of intelligent parking services can significantly reduce travelers’ cruising time for parking. Şengör et al. [15] proposed an energy management model for an EV parking lot (EVPL), based on real-time optimizations using linear programming, which can maximize the utilization of a parking lot. Dogaroglu and Caliskanelli [16,17] concluded that providing parking information, such as walking distance after parking and parking fees by the intelligent parking guidance system (IPGS), can decrease the individual driving distance and parking fees, balance the utilization of parking resources, and reduce pollutant emissions. Khaliq et al. [18] proposed a mutual authentication mechanism to solve certain privacy and security problems in existing intelligent parking systems.
As for intelligent parking recommendations, Huang et al. [19] introduced an intelligent decision support system for the purpose of guiding travelers toward available on-street parking spaces in urban areas. Li et al. [20] proposed a parking occupancy prediction method to optimize the use of parking spaces. Fu et al. [21] considered the types, position, walking distance, and other such factors of available parking spaces in a parking lot in order to design the optimal parking space recommendation model. The model could enable the parking spaces to be used effectively and thus reduce the parking search time. Shin and Jun [22] set fixed weights for the factors of travel time, walking distance after parking, parking price, and traffic congestion levels in order to recommend the most suitable available spaces in parking lots. The parking simulation showed that it could improve parking utilization. Based on the above studies, Safi et al. [23] added the factor of the driving safety level to the parking factors that were considered in the previous document; he found that the addition of this factor can help with effectively improving the utilization of parking resources and thus reduce fuel consumption. Shin et al. [24] proposed a parking recommendation model based on the neural network predictive control (NNPC) method. The simulation results showed that the model can alleviate traffic congestion and allow parking resources to be effectively utilized. Nine criteria for shared parking space allocations and parking route recommendations were proposed by Zhao et al. [25]; they also provided quantitative models for different conditions. The above studies mainly set certain weights for parking factors in their respective parking recommendation models, and they were found to improve the satisfaction of travelers with parking. While travelers usually have different preferences and psychological needs that affect their parking choice behaviors.
In terms of parking reservations, Sadreddini et al. [26] proposed an intelligent reservation system that considers the behavior, state-of-charge value, parking lot usage history of electric vehicle users, and parking space availability. The analytical hierarchy process in multi-criteria decision-making techniques was used in the intelligent reservation system to alleviate the parking difficulties of travelers’. Mei et al. (2019) [27] used genetic algorithms to simulate and optimize the reserve ratio of reserved parking spaces in a parking reservation system, and found that parking reservation can effectively improve the parking lot revenue while ensuring the interests of travelers. In order to reduce the number of inefficient reservation behaviors, Wang et al. [28] proposed a Blockchain-enabled Secure Framework for Energy-Efficient Smart Parking in Sustainable City Environment, the results of which show that it can effectively reduce the waste of parking resources. Liu et al. [29] designed a parking guidance system to recommend on-street parking spaces in real time for travelers to reserve spaces, based on the principle of the shortest distance. The simulation results showed that the system can significantly reduce travelers’ driving and walking costs. In order to minimize the total travel cost of all users, He et al. [30] proposed a real-time parking reservation service. They developed a mixed-integer planning model to efficiently allocate time slots and schedule drivers’ travel plans. Fu et al. [31] proposed a reservation-based parking recommendation model, which can filter parking lots based on travelers’ demand, and then recommend parking lots, based on the maximum utility of multiple attributes. The simulation experiment showed that the model can reduce cruising time for parking and parking costs, and improve the utilization of parking facilities.
Overall, for the studies on parking choice behavior, travel behavior characteristics, parking price, and walking distance after parking, it was found that these are all important factors that affect travelers’ parking choices. The relevant studies have analyzed the effectiveness regarding the application of intelligent parking service systems and discussed parking recommendations and reservations based on the parking factors. However, few studies have focused on parking choice behavior based on individual psychological factors of parking, nor have they considered the interests of travelers and parking managers for parking recommendations.
In this research, parking recommendation methods are studied based on travelers’ decision behaviors and psychological characteristics. The contents include the following: (1) Designing a sequential parking behavior survey and obtaining travelers’ decision-making behavior data at the decision points before a trip, during a trip, and near the destination. Then, a parking recommendation model and an individual parking decision process model were both established. (2) Based on a real-time acquisition of travelers’ psychological thresholds and levels of attention to parking factors, taking into account the benefits of travelers and managers, design static, and dynamic parking recommendation schemes. (3) Based on parking simulations, the effectiveness of different parking recommendation schemes was also explored. A deep analysis of the performance of the optimal scheme under different conditions was also carried out. Finally, certain suggestions are then put forward in order to solve parking problems.

3. Survey of Sequential Parking Decision Behavior under Parking Recommendations

3.1. Design and Implementation of the Survey

In order to analyze travelers’ parking decision behaviors under a parking recommendation system, a stated preference was used to design a parking behavior survey for shopping trips. First, we asked people if they drive regularly to filter out the groups that travel by car, and then we conducted a questionnaire. The survey contents consist of the following parts:
(1)
Personal information of the travelers.
The personal information includes gender, age, occupation, and monthly income.
(2)
Travelers’ attention to factors influencing parking choice.
According to the pilot survey, the influencing factors mainly include walking distance after parking, parking price, and driving time to the parking lot. The options for the degree of attention were set from “very unimportant (1)” to “very important (5)”, as per a 5-point Likert scale.
(3)
The individual psychological thresholds for the factors influencing parking choice.
The personal psychological threshold refers to the upper and lower limit of a person’s mental endurance or perceptual ability. In this article, when a parking factor in the parking lot is beyond or less than an individual’s acceptable value, the travelers will give up choosing the parking lot. For example, when the price of parking exceeds the acceptable value to the traveler, the traveler will abandon their choice of the parking lot. This paper obtains the threshold of the walking distance after parking, the threshold of the parking price, and the threshold of the available vacant parking spaces.
(4)
Stated preference survey on sequential parking decision behaviors.
It is assumed that a shopping travel scenario is presented by a screenshot of the Baidu map (Baidu Maps is an application that can display the location of the travel start and end points, travel distance, and travel routes, as well as the distribution of parking lots and parking information near travel destinations.). Establish shopping parking scenes for questionnaire survey. The trip origin of the home (i.e., the Century oriental garden community) and the trip destination of the Xidan Joy City are marked on the map. The travel distance is about 15 km and the travel activities of shopping last about 3 h. During the travel process, travelers can check the real-time parking information through the parking recommendation system and reserve a parking space as required.
The whole shopping travel process is divided into three decision points, including before the trip (15 km away from the destination), during the trip (7.5 km away from the destination), and near the destination (2 km away from the destination). Firstly, the traveler is asked whether they need to view the parking information before the trip. If their answer is “Yes”, then the real-time information of the parking lots within a certain area near the destination will be presented, including the distribution of parking lots, the number of vacant parking spaces, parking price, walking distance after parking, and travel time to the parking lot. Figure 1 shows an example of the distribution and real-time information of the parking lots. Then, the traveler is asked whether they need to make a decision about the parking lot based on the information presented. If their answer is “Yes”, they will be provided with an opportunity to choose a parking lot from among the available parking lots. At the same time, the parking recommendation system will provide a recommended parking lot to the traveler. The traveler can then reserve a parking lot of their own choice or as recommended by the parking recommendation system. If the traveler reserves a parking lot, the parking decision process will then end. Otherwise, the traveler will continue to answer the questions at the next decision point.
If the traveler does not view the parking information or reserve a parking lot at a decision point, they will be presented with the questions and the updated real-time parking information at the next decision point until they reach their destination. The setting of the questions for each decision point is similar. The whole parking decision process is shown in Figure 2. The recommended parking lot by the parking recommendation system is obtained through the parking recommendation model in Section 4.1. For travelers who do not view any of the parking information that is provided by the parking recommendation system during the whole travel process, they will then be provided with the choice of parking lots that have available parking spaces, ordered from the nearest to the furthest, after arriving at their destination.
The survey contents were compiled through the Questionnaire Star platform, which can perform the question-hopping design for different decision points during the travel process. The questionnaire was distributed through the internet from December 2020 to January 2021. Persons who own a car are qualified to answer the questionnaire. A total of 692 samples were returned, and there were 633 valid samples.

3.2. Graphic Characteristics and Differential Statistics of the Survey Data

(1)
Personal information.
The majority of participants in the survey samples were men, accounting for 61%. The age of the participants in the samples was mainly distributed between 26 and 35 years old, which accounted for 41%; this was followed by those who were 36~45 years old, which accounted for 33%. A total of 31% of the participants in the samples were professional and technical personnel, followed by 26% and 18% for freelancers and public institution/enterprise personnel, respectively. The participants’ personal monthly income mainly ranged from CNY 5000 to 10,000, accounting for 37%, followed by CNY 3000~5000 and CNY 10,000~15,000, accounting for 20% each, respectively.
(2)
Analysis of the travelers’ attention to the factors influencing parking choice.
As shown in Figure 3, most travelers think that the walking distance after parking, driving time, and parking price are the most important factors for their parking choice with respect to a shopping trip. The “more important” label accounts for the largest proportion of the three parking factors, accounting for 53%, 43%, and 38%, respectively. This, therefore, indicates that travelers pay more attention to the walking distance after parking, driving time, and parking price for their parking choice.
(3)
Analysis of travelers’ psychological thresholds for the parking factors influencing parking choice.
Figure 4 shows that 84% of travelers will not consider choosing a parking lot that has a walking distance of more than 800 m from the destination of their shopping trip. This indicates that travelers’ acceptable psychological threshold for the walking distance after parking is, in the main, less than 800 m. In addition, when the parking price for a parking lot exceeds 15 CNY/h, 86% of travelers will not consider the parking lot. The travelers’ psychological threshold for parking price was found to be, in the main, less than 15 CNY/h. A total of 88% of travelers would not consider choosing a parking lot with fewer than two available spaces, which means that the psychological threshold for available parking spaces is greater than or equal to two. The “Any” represents that they can accept any one choice.
(4)
Choice intention regarding the sequential parking decision process.
It can be seen from Figure 5 that the proportions of travelers who do not use the parking recommendation system to view information and those who view parking information without reserving a parking lot before a trip are 36% and 26%, respectively. Travelers who reserve a parking lot that is recommended by the parking reservation system and those who choose a parking lot on their own are 35% and 4%, respectively. Among the 62% of travelers who do not make a parking reservation before a trip, 42% of them do not use the parking recommendation system, and 13% of them view information without reserving a parking lot, respectively. In addition, only 7% of travelers choose to reserve a parking lot at this decision point. Overall, 55% of travelers who did not make parking reservations during the previous two decision points. As travelers drive closer to their destination, 34% of them still do not use the parking recommendation system, and 11% of them use the parking recommendation system only to view information without making a reservation. The proportion of travelers who reserve a parking lot is 10% at the decision point near the destination. It can be concluded that most travelers choose to make a parking reservation before the trip, accounting for 39%, which is then followed by those making a reservation when near the destination, and then those during the trip. Overall, 56% of travelers make parking reservations at some point during the whole travel process. Furthermore, most of them, accounting for 48%, choose to reserve the parking lot that is recommended by the recommendation system. This indicates that travelers are more willing than not to accept the parking recommendation system.

4. Parking Recommendation and Individual Parking Decision Process Models

4.1. Parking Recommendation Model

4.1.1. The Parking Recommendation Process, Considering Psychological Thresholds and Attention

It is assumed that in a parking recommendation system, the starting point and destination of travelers’ travel can be obtained through a mobile terminal application, and the attention and psychological threshold of parking factors that affect the parking choices can be obtained. At the same time, the parking recommendation system can receive real-time location information from travelers, and then collect real-time information from nearby parking lots to present it to them.
The parking recommendation process is as follows: Firstly, the candidate parking lots meeting the travelers’ needs at a decision point are tentatively selected based on the travel information, individual parking factors psychological threshold, and the real-time information of the parking lots near the destination. Secondly, the attribute information of the candidate parking lots is standardized, and the utility of each candidate parking lot is calculated from the perspective of travelers’ and managers’ interests, respectively. Finally, comprehensive utility scores of the candidate parking lots are obtained based on the adjustment coefficient for the proportions of the two-part utilities. Then, a better parking lot is obtained, based on the principle of utility maximization, and provided to the travelers. The recommendation process is divided into three steps, which is detailed below.
Step 1. The determination of candidate parking lots.
According to the information about parking lots at the decision point near the destination, firstly, the parking lots that meet the individual’s psychological threshold for the parking factors—specifically for all three factors, including walking distance after parking, parking price, and number of available parking spaces—are established. If there is no parking lot that satisfies all of the above screening criteria, select parking lots that meet the psychological threshold of any of the three factors as candidate parking lots. If there is still no parking lot that satisfies the above two screening criteria, all of the parking lots with vacant spaces within a certain range area around the destination can be used as candidate parking lots. The variable Z represents the set of candidate parking lots. Ztij is set to 1 if the parking lot j near the destination satisfies the screening process for traveler i at decision time t, otherwise it is 0.
Step 2. Standardization of the attribute information of candidate parking lots.
According to the candidate parking lots in Zti and their real-time information at a decision point, the influencing factor matrix Bti of candidate parking lots for traveler i at decision time t is obtained. In regard to Bti = {Lk, Ck, Otk}, k is the number of candidate parking lots, Lk is the walking distance from the kth parking lot to the destination, Ck is the parking price of the kth parking lot, and Otk is the number of available parking spaces in the kth parking lot at time t.
The influencing factors of the candidate parking lots are standardized from the perspective of travelers’ and managers’ interest, respectively. If the traveler’s benefit is considered, the standardization method for the factors of the candidate parking lots is shown in Formula (1). That is to say that the lower the parking price and the shorter the walking distance after parking, then the greater the traveler’s benefit. The standardized matrix for the influencing factors of the candidate parking lots for traveler i is R t i 1 = { l k , c k } .
l k = m i n ( L k ) L k c k = m i n ( C k ) C k ( k = 1 , 2 , m )
where lk and ck are the standardized values for the walking distance after parking and parking price for the kth candidate parking lot, respectively. In addition, m is the number of candidate parking lots.
From the perspective of parking managers’ benefits, the focus is more on the performance of parking lot utilization. In areas with high parking demand, using the parking recommendation system can guide travelers to park their car in parking lots that have more vacant spaces, thus reducing cruising for parking and improving the effective utilization of parking resources. Therefore, the standardization for the factor of the number of available parking spaces is shown in Formula (2). The standardized matrix for the influencing factors that relate to parking managers is R t g 2 = { o t k } .
o t k = O t k m a x ( O t k ) ( k = 1 , 2 , m )
where otk is the standardized value of the number of the available parking spaces of the kth candidate parking lot at time t.
Step 3. The calculation of the utility of candidate parking lots and parking recommendations.
Based on the traveler’s degree of attention to the influencing factors that are received in real time by the parking recommendation system, the weight vector Wi = [wil, wic] is the walking distance after parking and the parking price for traveler i is calculated by Formula (3).
w i q = d i q q = l c d i q   ( q = l , c )
where diq is the degree of attention to influencing factor q for traveler i.
The utilities of the candidate parking lots for traveler i at decision time t is obtained by Formula (4). Then, the parking lot with the highest utility is selected and recommended to traveler i.
E t i = λ W i R t i 1 + ( 1 λ ) R t g 2
where λ is the adjustment coefficient for the two-part utilities for travelers and managers.

4.1.2. Parking Recommendation Schemes That Consider the Interests of Travelers and Managers

The static and dynamic parking recommendation schemes can be obtained by the varying adjustment coefficients, as shown in Table 1. When the adjustment coefficient λ is set to fixed values of 1, 0, and 0.5, three static parking recommendation schemes can be obtained. Schemes 1, 2, and 3 show the cases that represent the maximization of travelers’ benefits, balancing parking resource utilization, and the combined benefits of travelers and parking managers, respectively. When the adjustment coefficient λ is dynamically changed with the utilization status of the parking lots, two dynamic parking recommendation Schemes 4 and 5, can be obtained. Scheme 5 includes a threshold to start parking regulation θ.

4.2. Individual Parking Decision Process Model

For the travelers who use the parking recommendation system to view the parking information, they may reserve their own chosen parking lot, or they may not make a reservation during the whole travel process. Their parking decision process is as follows: Firstly, based on the individual psychological thresholds regarding the parking factors, as well as the real-time information of the parking lot, the parking lots that satisfy travelers’ needs are screened as the initial acceptable parking lots. The screening process is similar to Step 1 in Section 4.1.1. If there is only one parking lot screened out, it will be the final parking choice for the trip during the individual decision process. If there is more than one parking lot selected through the above screening process, those remaining acceptable parking lots will be compared according to the traveler’s degree of attention to the parking factors of walking distance after parking, driving time, and parking price. From the most concerned to the least concerned factors, the comparison of the remaining parking lots will continue until only one parking lot is left. If no parking lot is obtained by filtering on all three parking factors, a randomly selected one from the initial acceptable parking lots is regarded as the final parking choice. For travelers who do not use the parking recommendation system, their parking decision process follows the principle of proximity parking.

5. Parking Simulation and Analysis in Different Parking Recommendation Schemes

5.1. Initial Settings of Parking Simulation

Based on the shopping travel scenario in the survey on sequential parking decision behaviors, the initial information of five parking lots within 1 km (after the preliminary investigation, it was found that the acceptable walking distance threshold for people after parking is mostly within 1 km) around the destination of Xidan Joy City is shown in Table 2. The number of parking spaces in each parking lot is about 200. Differentiated parking fees are applied in these parking lots and parking prices gradually decrease with the increasing distance from the destination.
It is assumed that car travelers are randomly generated in the range of 5–30 km around the shopping center. The trip generation rate follows a Poisson distribution with a mean value of 23 vehicles per 5 min. Travel distances of less than 5 km are mainly covered by non-motorized travel modes or public transport. We assume that people make the same proportion of parking choices before the trip, during the trip, and while near the destination as the data obtained in the questionnaire. According to the traffic status in Beijing, the average driving speed of cars is between 20 km/h and 50 km/h. The parking time distribution of car travelers in the shopping center is set according to the field survey. The parking time of less than 0.5 h accounts for 4%, 0.5 h to 1 h accounts for 8%, 1 h to 2 h accounts for 16%, 2 h to 3 h accounts for 32%, 3 h to 4 h accounts for 22%, and more than 4 h accounts for 18%. The vehicle will leave the parking lot automatically after reaching the parking time. The initial parked vehicles in the parking lot leave at a rate of three vehicles per 5 min.
The whole travel process is divided into three decision points, including before the trip (the location of trip generation), during the trip (midpoint of the trip), and near the destination (2 km away from the destination). Travelers can view parking information, and choose or reserve a parking lot at each decision point during the travel process. For the decision points before the trip, during the trip, and near the destination, the parking reservation proportions are 35%, 5%, and 8%, respectively, based on the parking decision behavior survey data in Figure 5. The travelers’ parking factor psychological thresholds and attentions to parking influencing factors are randomly assigned to the car travelers through the surveyed sample data. The parking reservation fee is CNY 2 per reservation. Parking is charged every 15 min, and anything less than 15 min is counted as 15 min.
Based on the parking recommendation model and individual parking decision process model, the travelers’ parking process for the shopping trip under the parking recommendation service is dynamically simulated using Python programming. The simulation lasts about 270 min. The threshold to start parking regulations is taken as 70%. As a whole, the number of vehicles arriving at the shopping center is greater than that which are leaving.

5.2. Analysis of Parking Simulation in Different Parking Recommendation Schemes

Parking simulation is carried out for the parking recommendation schemes from one to five to analyze the dynamic change in parking occupancy status, and to evaluate the performance of the schemes using multiple indicators.
(1)
The dynamic change in parking occupancy under different parking recommendation schemes.
As shown in Figure 6a, for the static parking recommendation Scheme 1, that is λ = 1, the parking recommendations are made from the perspective of maximizing travelers’ benefits. The parking lots that are closer to the destination are first recommended to travelers for their choice, and their occupancy rate increases quickly. When these nearby parking lots are full, the distant one is recommended for their parking choice and its occupancy rate begins to increase. Although the walking distance after parking for Park 5 is slightly farther than Park 4, the choice proportion of Park 5 was higher due to its lower parking price and higher integrated utility. For the static parking recommendation Scheme 2, that is λ = 0, parking recommendations are made from the perspective of manager’s benefits with the goal of balancing the utilization of parking resources. Figure 6b shows that the occupancy rates of the parking lots can basically be kept at the same level and present the same trends with the increase or decrease in parking demand after a period of adjustment. For the static Scheme 3, that is λ = 0.5, the interests of travelers and managers are considered at the same time for the purposes of parking recommendation. Figure 6c demonstrates that the changing trends in parking occupancy rates are the combined result of the parking recommendations of Scheme 1 and Scheme 2. Therefore, the smaller the adjustment coefficient λ, then the larger the proportion for the utility of the parking manager and the easier it is to achieve balanced utilization among the parking lots for a short period of time.
The dynamic parking recommendation Scheme 4 and Scheme 5 can change the recommendation mechanism according to the overall utilization of parking lots. As for Scheme 4, due to the relatively high mean value of the initial parking occupancy rate of 62% for these five parking lots, the parking recommendation is mainly made from the perspective of maximizing traveler benefits and balancing parking resource utilization. Figure 6d shows that the occupancy rates of parking lots near the destination increase along with the parking demand. In addition, the adjustment coefficient λ accordingly decreases gradually. As a result, the changing trend of the occupancy rate of each parking lot under Scheme 4 is essentially similar to that of Scheme 2. Parking recommendations under Scheme 5 are based on the median of the occupancy rates of parking lots, combined with the threshold in terms of starting parking regulations. Figure 6e shows that in the initial stage, the median occupancy rates of the parking lots near the destination is 62% and less than the threshold to start parking regulation of 70%. At this time, parking recommendations are made for the maximization of travelers’ benefits. Overall, the occupancy rate of parking lots gradually increases sequentially from the nearest to the farthest from the destination. When the simulation time approaches around 65 min, the median parking occupancy rates reaches the regulation threshold of 70%. It is then time to start the parking regulation by the parking recommendation system with the goal of balancing the utilization of parking resources while ensuring the interests of travelers. As a result, the trends in the occupancy rates of parking lots slowly moves closer to equilibrium over time.
(2)
Evaluation of simulation results in different parking recommendation schemes.
In order to evaluate the effects of the implementation of different parking recommendation schemes, multiple indicators are analyzed from the perspective of travelers and parking managers, as shown in Table 3. For the travelers’ benefits, the proportion of meeting the psychological threshold of the three parking factors is calculated, as well as the proportion of psychological thresholds that are primarily concerned with parking factors and the proportion of psychological thresholds that are secondarily concerned with parking factors. The three factors that affect the parking choice include the walking distance after parking, parking price, and the number of available parking spaces. Meanwhile, the evaluation indicators for travelers also consist of the average walking distance after parking and the average parking fees for their final chosen parking lot. The statistics are displayed in three groups. One is a parking reservation group, who make a reservation for a parking lot as recommended by the parking reservation system or by choosing on their own. The next is the non-parking reservation group, who do not make a reservation or use the parking recommendation system. Then, the last group includes all the travelers. The cumulative parking revenue is also given from the parking manager’s perspective.
Table 3 shows that there are about 90% or more travelers whose parking lot choice satisfies their psychological thresholds for all three factors as well as primary and secondary concerned factors under the parking recommended Schemes 2, 3, 4, and 5, while the satisfaction proportions for the parking recommendation of Scheme 1 are lower, which are 71%, 84%, and 82%, respectively. However, the average walking distance after parking for Scheme 1 is the shortest, with a value of 267 m/person. However, the average parking fee is approximately the same for all schemes. There are no significant differences for cumulative parking revenue among the schemes, with a maximum of CNY 28,961 in Scheme 1.
The satisfaction proportions of the psychological needs for the parking reservation group are higher than those for the non-parking reservation group under different parking recommendation schemes. Meanwhile, the average walking distance after parking for the parking reservation group is significantly higher than that of non-parking reservation group under the Schemes 2, 3, 4, and 5. The average parking fee for the non-parking reservation group was found to be slightly higher, which is because they do not view the parking information and mainly choose the closer parking lot to the destination with a higher parking price. This indicates that using the parking recommendation system can improve individual parking experience and satisfaction, although it may significantly increase the walking distance after parking.
In summary, the static parking recommendation Scheme 1, which recommends the parking lot from the perspective of travelers’ benefits, presents an imbalanced utilization of the parking lots due to the fact that the nearby parking lots are saturated while the distant ones are empty. The average walking distance after parking is relatively short, but the satisfaction proportion of psychological needs is also relatively low. Therefore, parking recommendation Scheme 1 is more suitable for the situations where the parking demand and the utilization of parking lots are relatively low.
For the recommended parking Schemes 2, 3, and 4, the changing trends of parking occupancy rates and the evaluation indicators are roughly the same. All these schemes can effectively balance the utilization of the parking lot resources. Meanwhile, the smaller the adjustment coefficient, the faster the speed of reaching the balanced parking utilization. Since the parking regulation by the parking recommendation system was launched from the start of the parking simulation, the average walking distance after parking is higher for the parking reservation group. The satisfaction proportions of the psychological thresholds of the parking factors were also relatively high. Therefore, Schemes 2, 3, and 4 were more suitable for situations where the parking demand is high and where the utilization of parking lots continues to be kept at a high level.
Dynamic parking recommendation Scheme 5 takes the interests of both travelers and parking managers into account. At the stage of low occupancy for the parking lots, the scheme can recommend the best parking lot for the maximization of the travelers’ benefits. When the parking demand increases and the occupancy of the parking lots reaches a high level, parking regulation is triggered by the parking recommendation system. The parking recommendation can make the utilization of parking resources more balanced and keep a certain proportion of available parking spaces for each parking lot, thus reducing cruising for parking. Based on the comprehensive analysis, the effect of parking recommendation Scheme 5 is the best, and thus further exploration of its performance under different conditions is conducted in the following chapters.

5.3. Analysis of Simulation Results in Different Initial Parking Utilization States

In order to analyze the effect of the dynamic parking recommendation Scheme 5 under different initial parking utilization states, two different travel scenarios are given based on the settings detailed in Section 5.1. One scenario represents the insufficient parking resources, whereby the initial number of available parking spaces for Park 1, Park 2, Park 3, Park 4, and Park 5 are set to 20, 30, 40, 50, and 80, respectively. The other scenario represents sufficient parking resources, whereby the initial number of available parking spaces for Park 1, Park 2, Park 3, Park 4, and Park 5 are set to 100, 120, 150, 160, and 180, respectively. The simulation is performed under these two scenarios to evaluate the results of the operation.
(1)
Dynamic change in parking occupancy in different initial parking utilization.
Compared with the changing trends of parking occupancy rates in Figure 6e, Figure 7a shows that the utilization of parking lot resources can be balanced quickly by parking recommendation Scheme 5 when there are a few vacant parking spaces, at first. The parking occupancy rate of each parking lot is close to an essentially balanced state at around 130 min. When there are sufficient initial parking spaces, as shown in Figure 7b, at the initial stage, parking recommendations are made mainly based on travelers’ utility maximization. Overall, the occupancy rates of the parking lots that are closer to the destination gradually increase and will be closer to saturation. At this time, the median of the parking occupancy rates of the parking lots reach the regulation threshold of 70% at about 135 min. Thus, the parking adjustment is then started from the perspective of balancing the utilization of parking resources while ensuring the interests of travelers. The farthest parking lot will then be recommended to travelers. Since implementing this, the occupancy rates of Park 4 and Park 5 have increased quickly.
(2)
Simulation results in different initial parking utilizations.
Table 4 shows that when the initial parking space is sufficient, the satisfaction ratio of the travelers to the psychological threshold of the three concerned parking factors, the satisfaction ratio of the psychological threshold of the primary parking concern factors, and the satisfaction ratio of the secondary parking concern factors are all slightly lower than those under the condition of insufficient initial parking space. The reason is that when there are sufficient parking spaces in the parking lot near the destination, recommendations are made with the goal of maximizing the benefits of travelers. In this situation, the parking lots that are closer to the destination are first recommended and then gradually saturated. Furthermore, subsequent travelers can only park their cars in the distant parking lot. Under the situation of scarce initial parking spaces, a balanced adjustment will be started at the beginning, which will not cause the over-saturated occupancy of the parking lots that are closer to the destination but will instead bring a higher average walking distance after parking.
The average parking fee and the cumulative parking revenue have no significant difference in the different initial parking utilization states. Compared with the non-parking reservation group, the satisfaction proportions of the psychological threshold for the parking factors of the parking reservation group are higher under different initial parking utilization states, while their average walking distance after parking is also higher. Therefore, parking recommendation Scheme 5 can be applied to different initial parking supply and demand situations and can play a good role in regulating parking demand.

5.4. Analysis of Simulation Results for Different Parking Reservation Proportions

According to the survey data, the parking reservation proportions are 35%, 5%, and 8% for the decision-making points of before the trip, during the trip, and near the shopping destination. These reservation proportions of the different decision points are changed in order to obtain two scenarios while keeping the other conditions unchanged. For reservation Scenario 1, the reservation proportions for the three sequential decision points are set to 5%, 35%, and 8%, respectively. For reservation Scenario 2, the reservation proportions are 8%, 5%, and 35%, respectively.
(1)
Dynamic change in the parking occupancy in different parking reservation proportions.
Under the initial reservation scenario, Figure 6e shows that the median parking occupancy rates reach the regulation threshold of 70% at around 65 min. At this time, the parking regulation is triggered by parking recommendation Scheme 5. It can be seen from Figure 8 that, while under reservation, Scenarios 1 and 2 possess higher reservation proportions during the trip and near the destination; however, the parking regulations have a late start at around 70 min and 75 min, respectively. At this moment, it is more difficult for the parking regulation to achieve a balanced utilization of parking lots because Park 1, Park 2, and Park 3 reach a relatively high and nearly saturated parking occupancy. When the reservation proportion before the trip is higher, the performance of balancing the utilization of parking resources in Figure 6e is better than that of reservation Scenarios 1 and 2. Therefore, the earlier the parking reservation is made during the whole travel process, the earlier the time of starting parking regulation and the easier it is to make the parking resources reach a balanced utilization.
(2)
Simulation results in different parking reservation proportions.
It can be seen from Table 5 that the satisfaction proportions of travelers’ psychological threshold for the parking factors for their chosen or reserved parking lots are higher under different parking reservation proportions, while keeping the whole parking reservations unchanged. At the initial setting (high proportion of pre travel parking reservations), the satisfaction proportions of the travelers’ psychological threshold for parking factors are slightly higher. Meanwhile, the average walking distance after parking is the shortest, at around 287 m/person for all travelers, 331 m/person for the parking reservation group, and 227 m/person for the non-parking reservation group, which was found to be lower than that of the reservation Scenarios 1 and 2. In addition, the cumulative parking revenue for managers was also found to be higher. There was no significant difference found in the average parking fee under different parking reservation scenarios.
In general, the parking recommendation Scheme 5 can achieve the effect of balancing the utilization of parking resources according to parking demand and thus can present a high satisfaction proportion of travelers’ psychological thresholds for parking factors under different parking reservation proportions at different stages of the travel process. Therefore, making parking reservations earlier, the system regulates the start time earlier, and the balance of parking lot resources is better. Meanwhile, the average walking distance after parking was lower and the parking revenue was higher. Therefore, providing sufficient parking information to promote travelers’ parking reservations before the trip can effectively reduce the cruising for parking and thus enhance the utilization of parking resources.

5.5. Analysis of Simulation Results in Different Parking Regulation Thresholds

In order to analyze the impact of regulation thresholds on the performance of parking utilization, two changed thresholds of 50% and 90% were given based on the initial setting of 70%.
(1)
Dynamic change in parking occupancy in different parking regulation thresholds.
Compared with the performance under the parking regulation threshold of 70% in Figure 6e, Figure 9a shows that the parking regulation is triggered in order to balance the resource utilization at the start due to the higher occupancy rates of the parking lots under the lower regulation threshold of 50%. Additionally, the occupancy rates in parking lots reach a basically balanced state at around 210 min. When the parking regulation threshold is set to 90%, as shown in Figure 9b, the time to start the parking regulations is delayed to 90 min, compared with the initial setting of 70%. Due to the high threshold that is required to start parking regulations, certain parking lots are thus already nearly saturated, and the parking regulation is thus frequently started at different periods of time. When the median parking occupancy rates of the parking lots are less than 90% again, the parking recommendation will be made to maximize the benefits of travelers, which leads to an increase in the occupancy of car parks close to peoples destinations. Eventually, it is difficult for the adjustment to make parking resource utilizations reach the overall equilibrium state.
(2)
Simulation results in different parking regulation thresholds.
Table 6 shows that when the threshold to start parking regulation changes from 50% to 90%, then the satisfaction proportions of travelers’ psychological threshold for parking factors gradually decrease. Meanwhile, the average walking distance after parking for all travelers and the parking reservation group also decreases; moreover, the cumulative parking revenue thus increases. Meanwhile, the satisfaction proportions of the travelers’ psychological threshold for parking factors for the parking reservation group were higher than those for the non-parking reservation group under different parking regulation thresholds. However, the average walking distance for the parking reservation group was found to be relatively longer.
Overall, the higher the regulation threshold is, the later the time to start the parking regulation and the more difficult it is to achieve a balanced utilization of parking lots. When the threshold to start parking regulation is set at 70%, the satisfaction proportions of the travelers’ psychological threshold for parking factors and the cumulative parking revenue are higher. In addition, the average walking distance after parking is also relatively shorter. Therefore, the threshold to start parking regulation should be set neither too high nor too low, both from the perspective of improving travelers’ travel experiences and also the utilization of parking resources.

6. Conclusions

Intelligent parking services can provide travelers with real-time parking information, as well as help reduce cruising time and balance the utilization of parking resources by using parking recommendations and reservations. It is an effective method for solving the parking problems in big cities. Based on a parking survey, this research analyzed sequential parking decision behaviors at different stages of the travel process. Then, the parking recommendation model and the individual parking decision model, which considered travelers’ psychological characteristics, were established in order to analyze the applicability of different parking recommendation schemes. Finally, the optimal dynamic parking recommendation scheme is selected and its applicability in different situations is explored. The conclusions are as follows:
Based on the survey of sequential parking decision behavior, it was found that travelers pay more attention to walking distance after parking, driving time, and parking price when making their parking choice. Travelers’ acceptable psychological thresholds for parking factors with respect to their parking lot choice were found to be, in the main, less than 800 m for walking distance after parking, less than 15 CNY/h for parking price, and more than two for available parking spaces. In general, travelers were more willing to accept and use the parking recommendation system. The proportion of parking reservations was 56%, with the highest value of 39% at the stage before the trip, followed by near the destination, and then during the trip. Most travelers, accounting for 48%, reserve the parking lot recommended by the recommendation system.
The parking recommendation model and the individual parking decision model, considering the travelers’ psychological thresholds and attention to parking factors—which were received in real time—were established. The static and dynamic parking recommendation schemes, based on a changing adjustment coefficient, were designed. The simulation results show that the satisfaction proportion of the psychological threshold for parking factors for the parking reservation group was found to be, under different parking recommendation schemes, higher than that for the non-parking reservation group. Using the parking recommendation system, individual parking experiences and satisfaction can be improved, although the system may slightly increase the walking distance after parking.
The static parking recommendation Scheme 1 presents the phenomenon of an unbalanced utilization of parking spaces for nearby and distant parking lots. Since it recommends the parking lot from the perspective of maximizing travelers’ benefits, the average walking distance after parking was relatively short, but the satisfaction proportions of psychological needs for the parking factors were relatively low. Scheme 1 is more suitable for situations where the utilization of parking lots is relatively low. The recommended parking Schemes 2, 3, and 4 have similar, better operational performances in terms of balancing the utilization of parking resources. The satisfaction proportions of the psychological needs for parking factors are relatively high, while the average walking distance after parking is also higher. These schemes are more suitable for situations where the utilization of parking lots is kept at a high level. The dynamic parking recommendation Scheme 5 can take the interests of both travelers and parking managers into account. At a stage with a low occupancy rate for parking lots, this scheme can recommend the parking lot from the perspective of maximizing travelers’ benefits. When the occupancy of the parking lots is high enough, parking regulations will be triggered to make parking resource utilization more balanced. Therefore, parking recommendation Scheme 5 is better.
Based on the simulation of parking recommendation Scheme 5 under different conditions, it is concluded that dynamic parking recommendation Scheme 5 can quickly balance the utilization of parking resources in the case of a shortage of parking spaces. When the initial parking space is sufficient, parking recommendations can be carried out, at first, based on the benefits to travelers. Then, the parking regulation can start when the parking occupancy reaches a high level. Parking recommendation Scheme 5 can be applied to different parking utilization conditions and plays a good role in regulating parking demand distribution. Making parking reservations before the trip has a better effect in terms of balancing the utilization of parking resources. Meanwhile, the average walking distance after parking is lower and the parking revenue is higher. Therefore, providing sufficient parking information to promote travelers’ parking reservations before the trip is beneficial for reducing cruising for parking and in terms of optimizing parking resource utilization. For the dynamic parking recommendation, the threshold to start parking regulation should be set neither too high nor too low from the perspectives of increasing travelers’ parking experience and in balancing the utilization of parking resources.
This research explored parking recommendation methods considering travelers’ psychological characteristics. This article also has some limitations. For example, only shopping travel scenarios are discussed. In the future, we should explore the applicability of more parking and travel scenarios, such as office or work trips. Another limitation is that the simulations in this paper assume that people’s parking choices are the same as those obtained from the questionnaire. We should investigate and obtain more parking reservation data under travel scenarios. The survey data can be enriched to improve the decision models for parking choice and reservation for different groups during the travel process; in other words, it will be better to verify the effects of implementing the parking recommendation models and schemes through practical examples.

Author Contributions

Methodology, data curation, and writing original draft, N.X.; conceptualization, funding acquisition, and writing—review and editing, H.Q.; supervision, Y.Z., Q.P. and Z.L. All authors have read and agreed to the published version of the manuscript.

Funding

This research was supported by the Beijing Natural Science Foundation (8212002) and the National Natural Science Foundation (71971005).

Institutional Review Board Statement

Ethical review and approval were waived for this study, as no personal identity was involved or reported.

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Data Availability Statement

The data used to support the findings of this study are available from the corresponding authors upon request.

Acknowledgments

The authors would like to thank anonymous reviewers for their valuable comments and suggestions.

Conflicts of Interest

The authors declare that there are no conflicts of interest regarding the publication of this paper.

References

  1. Shoup, D.C. Cruising for parking. Transp. Policy 2006, 13, 479–486. [Google Scholar] [CrossRef]
  2. Zhu, Y.; Ye, X.; Chen, J.; Yan, X.; Wang, T. Impact of Cruising for Parking on Travel Time of Traffic Flow. Sustainability 2020, 12, 3079. [Google Scholar] [CrossRef]
  3. Giuffrè, T.; Siniscalchi, S.; Tesoriere, G. A novel architecture of parking management for smart cities. Procedia–Soc. Behav. 2012, 53, 16–28. [Google Scholar] [CrossRef]
  4. Caliskan, M.; Barthels, A.; Scheuermann, B.; Mauve, M. Predicting Parking Lot Occupancy in Vehicular Ad Hoc Networks. In Proceedings of the IEEE 65th Conference on Vehicular Technology, Dublin, Ireland, 22–25 April 2007. [Google Scholar]
  5. Paul, B. On-Street Parking Management; Deutsche Gesellschaft Für International Zusammenarbeit: Berlin, Germany, 2016. [Google Scholar]
  6. Arnott, R.; Inci, E. An integrated model of downtown parking and traffic congestion. J. Urban Econ. 2005, 60, 418–442. [Google Scholar] [CrossRef]
  7. Van der Waerden, P.; Timmermans, H.; da Silva, A.N.R. The influence of personal and trip characteristics on habitual parking behavior. Case Stud. Transp. Policy 2015, 3, 33–36. [Google Scholar] [CrossRef]
  8. Zhang, R.; Zhu, L. Curbside parking pricing in a city centre using a threshold. Transp. Policy 2016, 52, 16–27. [Google Scholar] [CrossRef]
  9. Khaliq, A.; van der Waerden, P.; Janssens, D.; Wets, G. A conceptual framework for forecasting car driver’s on-street parking decisions. Transp. Res. Procedia 2019, 37, 131–138. [Google Scholar] [CrossRef]
  10. Soto, J.J.; Márquez, L.; Macea, L.F. Accounting for attitudes on parking choice: An integrated choice and latent variable approach. Transp. Res. Part A Policy Pract. 2018, 111, 65–77. [Google Scholar] [CrossRef]
  11. Tian, Q.; Yang, L.; Wang, C.L.; Huang, H.-J. Dynamic pricing for reservation-based parking system: A revenue management method. Transp. Policy 2018, 71, 36–44. [Google Scholar] [CrossRef]
  12. Lei, C.; Ouyang, Y.F. Dynamic pricing and reservation for intelligent urban parking management. Transp. Res. Part C Emerg. Technol. 2017, 77, 226–244. [Google Scholar] [CrossRef]
  13. Fahim, A.; Hasan, M.; Chowdhury, M.A. Smart parking systems: Comprehensive review based on various aspects. Heliyon 2021, 7, e07050. [Google Scholar] [CrossRef] [PubMed]
  14. Cao, J.; Menendez, M. Quantification of potential cruising time savings through intelligent parking services. Transp. Res. Part A Policy Pract. 2018, 116, 151–165. [Google Scholar] [CrossRef]
  15. Şengör, İ.; Güner, S.; Erdinç, O. Real-time algorithm based intelligent EV parking lot charging management strategy providing PLL type demand response program. IEEE Trans. Sustain. Energy 2020, 12, 1256–1264. [Google Scholar] [CrossRef]
  16. Dogaroglu, B.; Caliskanelli, S.P. Investigation of car park preference by intelligent system guidance. Res. Transp. Bus. Manag. 2020, 37, 100567. [Google Scholar] [CrossRef]
  17. Dogaroglu, B.; Caliskanelli, S.P.; Tanyel, S. Comparison of intelligent parking guidance system and conventional system with regard to capacity utilization. Sustain. Cities Soc. 2021, 74, 103152. [Google Scholar] [CrossRef]
  18. Khaliq, A.A.; Anjum, A.; Ajmal, A.B.; Webber, J.L.; Mehbodniya, A.; Khan, S. A secure and privacy preserved parking recommender system using elliptic curve cryptography and local differential privacy. IEEE Access 2022, 10, 56410–56426. [Google Scholar] [CrossRef]
  19. Huang, Y.H.; Hsieh, C.H. A decision support system for available parking slots on the roadsides in urban areas. Expert Syst. Appl. 2022, 205, 117668. [Google Scholar] [CrossRef]
  20. Li, J.; Qu, H.; You, L. An Integrated Approach for the Near Real-Time Parking Occupancy Prediction. IEEE Trans. Intell. Transp. Syst. 2022, 24, 3769–3778. [Google Scholar] [CrossRef]
  21. Fu, J.; Chen, Z.; Sun, R. Research on intelligent terminal oriented optimal parking space recommendation model. In Proceedings of the 17th International IEEE Conference on Intelligent Transportation Systems (ITSC), Qingdao, China, 8–11 October 2014; pp. 2373–2378. [Google Scholar]
  22. Shin, J.H.; Jun, H.B. A study on smart parking guidance algorithm. Transp. Res. Part C Emerg. Technol. 2014, 44, 299–317. [Google Scholar] [CrossRef]
  23. Safi, Q.G.K.; Luo, S.L.; Pan, L.M.; Liu, W.; Hussain, R.; Bouk, S.H. SVPS: Cloud-based smart vehicle parking system over ubiquitous VANETs. Comput. Netw. 2018, 138, 18–30. [Google Scholar] [CrossRef]
  24. Shin, J.H.; Jun, H.B.; Kim, J.G. Dynamic control of intelligent parking guidance using neural network predictive control. Comput. Ind. Eng. 2018, 120, 15–30. [Google Scholar] [CrossRef]
  25. Zhao, P.; Wei, L.; Pan, D.; Dong, S. Multicriteria Model for Shared Parking and Parking Route Recommender Systems. J. Adv. Transp. 2022, 2022, 8369362. [Google Scholar] [CrossRef]
  26. Sadreddini, Z.; Guner, S.; Erdinc, O. Design of a decision-based multicriteria reservation system for the EV parking lot. IEEE Trans. Transp. Electrif. 2021, 7, 2429–2438. [Google Scholar] [CrossRef]
  27. Mei, Z.Y.; Zhang, W.; Zhang, L.H.; Wang, D. Optimization of reservation parking space configurations in city centers through an agent-based simulation. Simul. Model. Pract. Theory 2020, 99, 102020. [Google Scholar] [CrossRef]
  28. Wang, J.; Zhu, C.; Miao, C.; Zhu, R.; Zhang, X.; Tang, Y.; Huang, H.; Gao, C. BPR: Blockchain-Enabled Efficient and Secure Parking Reservation Framework with Block Size Dynamic Adjustment Method. IEEE Trans. Intell. Transp. Syst. 2022, 24, 3555–3570. [Google Scholar] [CrossRef]
  29. Liu, K.S.; Gao, J.; Wu, X.; Lin, S. On-street parking guidance with real-time sensing data for smart cities. In Proceedings of the 2018 15th Annual IEEE International Conference on Sensing, Communication, and Networking (SECON), Hong Kong, China, 11–13 June 2018; pp. 1–9. [Google Scholar]
  30. He, H.; Zhang, Z.; Yan, P. A real-time reservation service for smart parking system. In Proceedings of the 2018 15th international conference on service systems and service management (ICSSSM), Hangzhou, China, 21–22 July 2018; pp. 1–6. [Google Scholar]
  31. Fu, J.; Chen, Z.; Sun, R.; Yang, B. Reservation based optimal parking lot recommendation model in Internet of Vehicle environment. China Commun. 2014, 11, 38–48. [Google Scholar] [CrossRef]
Figure 1. Parking lot distribution and real-time information around the trip destination.
Figure 1. Parking lot distribution and real-time information around the trip destination.
Sustainability 15 06808 g001
Figure 2. The design of a parking decision process for different decision points.
Figure 2. The design of a parking decision process for different decision points.
Sustainability 15 06808 g002
Figure 3. Attention to factors influencing parking choice.
Figure 3. Attention to factors influencing parking choice.
Sustainability 15 06808 g003
Figure 4. Psychological thresholds for the parking factors influencing parking choice.
Figure 4. Psychological thresholds for the parking factors influencing parking choice.
Sustainability 15 06808 g004
Figure 5. Parking choice proportions at different travel decision points.
Figure 5. Parking choice proportions at different travel decision points.
Sustainability 15 06808 g005
Figure 6. Dynamic change in the parking occupancy rate for different parking recommendation schemes: (a) Scheme 1; (b) Scheme 2; (c) Scheme 3; (d) Scheme 4; and (e) Scheme 5.
Figure 6. Dynamic change in the parking occupancy rate for different parking recommendation schemes: (a) Scheme 1; (b) Scheme 2; (c) Scheme 3; (d) Scheme 4; and (e) Scheme 5.
Sustainability 15 06808 g006
Figure 7. The dynamic change in parking occupancy rates for different initial parking utilizations: (a) insufficient scarce parking spaces and (b) sufficient parking spaces.
Figure 7. The dynamic change in parking occupancy rates for different initial parking utilizations: (a) insufficient scarce parking spaces and (b) sufficient parking spaces.
Sustainability 15 06808 g007
Figure 8. The dynamic change in parking occupancy rates for different parking reservation proportions: (a) reservation Scenario 1; (b) reservation Scenario 2.
Figure 8. The dynamic change in parking occupancy rates for different parking reservation proportions: (a) reservation Scenario 1; (b) reservation Scenario 2.
Sustainability 15 06808 g008
Figure 9. The dynamic change in parking occupancy rates for different parking regulation thresholds: (a) parking regulation threshold of 50% and (b) parking regulation threshold of 90%.
Figure 9. The dynamic change in parking occupancy rates for different parking regulation thresholds: (a) parking regulation threshold of 50% and (b) parking regulation threshold of 90%.
Sustainability 15 06808 g009
Table 1. Parking recommendation schemes.
Table 1. Parking recommendation schemes.
SchemesAdjustment
Coefficient λ
Description
Static parking recommendationScheme 1 λ = 1 Parking recommended with the goal of maximizing the benefits of travelers.
Scheme 2 λ = 0 Parking recommended with the goal of balancing parking resource utilization.
Scheme 3 λ = 0.5 Parking recommended based on the combined benefits of travelers and parking managers.
Dynamic parking recommendationScheme 4 λ = 1 P t k ¯ P t k ¯ is the average occupancy rate of the parking lots near the destination at time t. If P t k ¯ is high, then the parking lot is recommended mainly based on the benefits of parking managers. If P t k ¯ is low, then the parking recommendation is made mainly based on maximizing the benefits of travelers.
Scheme 5 λ = 1 P t k ¯ ,
if   P t k m > θ ;
λ = 1 ,
if   P t k m θ
P t k m is the median of the occupancy rates of the parking lots near the destination at time t. When P t k m is greater than θ , the parking regulation is triggered, and the parking recommendation is made based on the benefits of parking managers. When P t k m is less than or equal to θ , the parking lot is recommended with the goal of maximizing the benefits of the travelers.
Table 2. Initial information of the parking lots near the destination.
Table 2. Initial information of the parking lots near the destination.
Parking
Lots
Distance from the Parking Lot to the Destination,
the Joy City (m)
The Number of
Available Parking Spaces
Parking
Price (CNY/h)
P1602010
P21008010
P3250808
P45201008
P56001005
Table 3. The simulation results in the different parking recommendation schemes.
Table 3. The simulation results in the different parking recommendation schemes.
SchemesIndicators Related to Travelers
(All/Parking Reservation Group/Non-Parking Reservation Group)
Indicators Related to Parking Manager
Satisfaction of Psychological Needs (%)Average Walking
Distance after Parking (m/Person)
Average Parking Fee (Including Reservation Fee)
(CNY/Person)
Cumulative
Parking Revenue (CNY)
All Three
Factors
Primary Concerned FactorSecondary Concerned Factor
Scheme 171/74/6684/85/8382/84/80267/262/27324/24/2428,961
Scheme 297/98/9692/95/8991/92/89298/362/21124/23/2328,264
Scheme 397/98/9692/95/8991/92/89294/359/20924/23/2428,309
Scheme 497/98/9692/95/8991/92/89297/362/20824/23/2528,239
Scheme 592/92/9191/92/8789/90/88287/331/22724/23/2428,499
Table 4. Simulation results in different initial parking utilizations.
Table 4. Simulation results in different initial parking utilizations.
ScenarioIndicators Related to Travelers
(All/Parking Reservation Group/Non-Parking Reservation Group)
Indicators Related to Parking Manager
Satisfaction of Psychological Needs (%)Average Walking
Distance after Parking
(m/Person)
Average Parking Fee (Including Reservation Fee)
(CNY/Person)
Cumulative
Parking Revenue
(CNY)
All Three
Factors
Primary Concerned FactorsSecondary Concerned Factors
Initial
settings
92/92/9191/92/8789/90/88287/331/22724/23/2428,499
Insufficient parking spaces93/94/9191/94/8790/91/88300/364/21324/23/2428,170
Sufficient parking spaces87/90/8489/91/8788/89/86261/276/23924/25/2429,034
Table 5. Simulation results in different parking reservation proportions.
Table 5. Simulation results in different parking reservation proportions.
ScenarioIndicators Related to Travelers
(All/Parking Reservation Group/Non-Parking Reservation Group)
Indicators Related to Parking Manager
Satisfaction of Psychological Needs (%)Average Walking
Distance after Parking
(m/Person)
Average Parking Fee (Including Reservation Fee)
(CNY/Person)
Cumulative
Parking Revenue
(CNY)
All Three
Factors
Primary Concerned FactorsSecondary Concerned Factors
Initial settings92/92/9191/92/8789/90/88287/331/22724/23/2428,499
Reservation Scenario 190/92/8790/93/8788/90/88296/342/23524/23/2427,880
Reservation Scenario 289/92/8589/93/8588/90/86292/340/23124/23/2427,513
Table 6. Simulation results in different parking regulation thresholds.
Table 6. Simulation results in different parking regulation thresholds.
Threshold to Start Parking RegulationIndicators Related to Travelers
(All/Parking Reservation Group/Non-Parking Reservation Group)
Indicators Related to Parking Manager
Satisfaction of Psychological Needs (%)Average Walking
Distance after Parking
(m/Person)
Average Parking Fee (Including Reservation Fee)
(CNY/Person)
Cumulative
Parking Revenue (CNY)
All Three
Factors
Primary Concerned FactorsSecondary Concerned Factors
Initial setting of 70%92/92/9191/92/8789/90/88287/331/22724/23/2428,499
50%97/98/9692/95/8991/92/89297/362/20824/22/2528,239
90%85/88/8188/91/8487/89/84276/307/23524/24/2428,679
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Qin, H.; Xu, N.; Zhang, Y.; Pang, Q.; Lu, Z. Research on Parking Recommendation Methods Considering Travelers’ Decision Behaviors and Psychological Characteristics. Sustainability 2023, 15, 6808. https://doi.org/10.3390/su15086808

AMA Style

Qin H, Xu N, Zhang Y, Pang Q, Lu Z. Research on Parking Recommendation Methods Considering Travelers’ Decision Behaviors and Psychological Characteristics. Sustainability. 2023; 15(8):6808. https://doi.org/10.3390/su15086808

Chicago/Turabian Style

Qin, Huanmei, Ning Xu, Yonghuan Zhang, Qianqian Pang, and Zhaolin Lu. 2023. "Research on Parking Recommendation Methods Considering Travelers’ Decision Behaviors and Psychological Characteristics" Sustainability 15, no. 8: 6808. https://doi.org/10.3390/su15086808

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop