Digital Transformation and Sustainability in Perishable Product Logistics: Emerging Themes and Future Directions in the Industry 5.0 Context Through a Systematic Literature Review

Abstract

Sustainable logistics for perishable products has received heightened attention, as evidenced by an expanding corpus of academic research. Aligning with this international perspective, the current study provides a comprehensive assessment of the research landscape, identifying cutting-edge developments and emerging trends in sustainable logistics for perishable products in the context of the transition from Industry 4.0 toward Industry 5.0. Utilizing the Systematic Reviews and Meta-Analyses (PRISMA) framework, a systematic review is conducted on 104 peer-reviewed articles sourced from the Scopus and Web of Science databases, published between 2021 and 2025. Based on an in-depth examination of seven key research themes, various research directions are suggested to guide future investigations. The observations and conclusions drawn from this analysis aim to establish a solid foundation for advancing knowledge on digitalized and sustainable logistics for perishable goods.

1. Introduction

In recent years, increasing public concern regarding environmental and social responsibility has heightened attention on the performance of perishable logistics and its associated impacts [1]. As a result, research on sustainable perishable logistics has provided a comprehensive understanding of the synergies among its economic, environmental, and social dimensions. The sustainability of perishable logistics is influenced by two interconnected and cascading factors: the inherent perishability of the products, and their specific transportation and storage requirements. Perishable products are typically categorized into four distinct groups based on their decay characteristics: (1) those with a non-strict fixed lifetime (e.g., foods, fruits, vegetables, and flowers); (2) those with a strict fixed lifetime (e.g., blood products and pharmaceuticals); (3) those with a random shelf life (e.g., gasoline and radioactive substances); and (4) those that gradually deteriorate over time (e.g., alcohol and pesticides) [2]. Given the perishable nature of the products, the associated supply chains face persistent challenges in many major industries, such as the food [3], agri-food [4], and pharmaceutical and medical industries [5,6]. Such supply chains are intrinsically complex due to the requirements of on-time delivery while keeping product quality, safety, and freshness. Any failure in these supply chains can lead to significant economic losses, waste of valuable resources, and even health risks for consumers. Between one-third and one-quarter of the food produced for human consumption was lost or wasted in 2014 [7]. Each year in Iran, food waste is estimated at around 15 billion dollars, with the majority resulting from spoilage [8]. More critically, in the Republic of Serbia, approximately 247,000 tons of food are wasted each year, largely due to inadequate inventory management and logistics practices [9]. On the other hand, the transportation of perishable goods exhibits distinct characteristics compared to that of non-perishable goods, primarily due to the time sensitivity of these products. Consequently, the requirement to maintain product freshness, minimize spoilage, and meet critical operational requirements in perishable goods logistics increases both costs and environmental impacts. For instance, cold chain logistics is characterized by energy intensities substantially higher than those of conventional logistics systems, primarily driven by refrigeration and thermal management, such as temperature control and insulated transport [10]. These specific requirements have further exacerbated the environmental issues through multiple threats, such as greenhouse gas emissions and the leakage of refrigerants (e.g., hydrofluorocarbons) used in cooling systems, as well as carbon dioxide (CO₂) emitted from transportation activities, all of which significantly contribute to global warming potential [11]. According to the European Commission (https://environment.ec.europa.eu/news/field-fork-global-food-miles-generate-nearly-20-all-co2-emissions-food-2023-01-25_en, accessed on 14 December 2025), transportation and logistics activities related to perishable products accounted for approximately 20% of global CO₂ emissions in 2023. When combined with the environmental impacts of product waste, this poses a significant challenge not only to sustainability but also to transparency and ethical sourcing within perishable supply chains. Therefore, perishable logistics and efforts to enhance their sustainability have become a critical research focus in recent decades, particularly under the growing pressure to achieve the 2030 Sustainable Development Goals (SDGs) [12].

Like other industries, technology plays a critical role in perishable goods logistics, with diverse technologies integrated for risk mitigation, efficiency improvements, and complexity management, especially with respect to freshness preservation and safety assurance. These advancements help reduce economic losses, increase customer satisfaction, and promote sustainability. The adoption of digital technologies has been widely recognized as a key enabler for improving efficiency through monitoring various stages of the supply chain [13], and sustainability in agri-food processing [14]. The growing adoption of these technologies has motivated a substantial body of research examining Industry 4.0 (I4.0) technologies, such as artificial intelligence (AI), the Internet of Things (IoT), robotics, and cloud computing, and their applications within perishable product logistics and supply chains. These technologies are widely acknowledged for their ability to improve operational efficiency, facilitate real-time monitoring, and support data-driven decision-making. Although they have significantly enhanced the efficiency of perishable logistics, their primary focus has been on enabling mass production through smart factories [15,16].

Amid the ongoing technological revolution, the transition from I4.0 to the emerging Industry 5.0 (I5.0) paradigm has gained prominence in the 2020s [17]. While I4.0 focused primarily on techno-centric automation and production efficiency [18], I5.0 heralds a transformative shift toward a more holistic and value-centric approach. Unlike I4.0’s efficiency-driven focus, I5.0 reorients the same technologies toward sustainability, human-centricity, and systemic resilience [19], offering perishable logistics not merely as a tool for cost reduction but an enabler for broader socio-environmental transformation. For instance, I5.0 leverages AI and machine learning to support stakeholder decision-making enabling market-driven predictions that minimize food spoilage and waste [20]; autonomous delivery systems and drones perform request-responsive last-mile and door-to-door logistics [21]; IoT and 5G technologies are integrated into service platforms to mitigate environmental change risks [22]; and IoT, blockchain, and edge computing ensure data security and supply chain sustainability in cold chain logistics [23].

Despite the growing transition from I4.0 to I5.0, its adoption and particularly its application in the realm of perishable logistics remain at an initial stage and are still largely underexplored [24]. As a result, there is an urgent need to understand the enablers of I5.0 adoption and its subsequent influence on the sustainability of perishable supply chains. To address this, the current systematic review attempts to provide a comprehensive overview of existing research on sustainability-oriented perishable product logistics within the emerging I5.0 landscape. This analysis highlights the latest advancements in the field, uncovers critical gaps in the current body of literature, and outlines a strategic roadmap for future investigations.

1.1. Relevant Existing Review Papers and the Contribution of the Present Paper

The findings from an extensive literature survey indicate that most existing review papers tend to focus exclusively on either technological integration or sustainability efforts for specific perishable categories, rather than both. Several reviews have examined the evolution of fresh fruit and vegetable supply chains over 20 years [25]; decision-making tools in food cold chain logistics [26,27]; and sustainability drivers [28,29]. From the technological perspective, Haji et al. [15] identified how various technologies enhance the quality of perishable items throughout the supply chain process. Rejeb et al. [30] provided an overview of digital transformation in the food supply chain. Yadav et al. [31] examined the application of I4.0 technologies in the supply chain for agricultural products. These reviews primarily focus on specific product types, particularly food items.

It is evident that the interplay of technology applications and sustainability enhancements of perishable logistics has received relatively limited attention in the existing review studies. Only a few recent works have explored the relationship between the two aspects in these fields. One such study is by Remondino and Zanin [32], who extensively reviewed the impact of digitalization and sustainability in the cold chain 4.0 for the agri-food sector. Through a systematic literature review, Zhao et al. [33] analyzed the factors driving the implementation of I4.0 tools to improve the sustainability of agricultural food supply chains. This concept was further explored by Ertz et al. [34], who examined the integration of digital and sustainable technologies within the cold chain 4.0. Nevertheless, the existing studies are largely restricted to applications of I4.0. The possibilities offered by the I5.0 framework have been investigated by Singh et al. [35] and Math et al. [36], focusing respectively on the food sector and the healthcare supply chain. Despite being among the most relevant and pioneering contributions, the scopes of these two review papers remain relatively broad and pay limited attention to food and healthcare products with perishable natures. The key contributions of the reviewed papers are summarized in

Table 1. Recent review studies on sustainable logistics and supply chain of perishable products.

Authors	Aims	No. of. Papers	Time Period	Review Type	Notes
Tort et al. [25]	To review existing studies on fresh fruit and vegetable supply chains.	118	2000–2020	SLR	Mainly focused on fresh fruit and vegetables
Zhang and Mohammad [26]	To assess the literature on perishable food cold chain logistics, emphasizing decision-making tools for sustainability and the role of smart technologies.	80 (only WoS)	2010–2023	SLR	Only focused on cold foods with limited attention on the application of technologies
Khalid et al. [27]	To review and assess the existing literature on food cold chains in relation to risk management and supply chain sustainability principles.	155	2000–2023	LR	Focused on food cold chain management
Shetty et al. [28]	To review the emerging research trends of sustainability research within perishable food supply chains.	389	2009–2023	SLR	Limited to the perishable food supply chain, with a brief review of some applications of IoT and blockchain
de Castro Moura Duarte et al. [29]	To review and identify the key factors affecting the sustainability of fresh food supply chains.	39	2007–2022	SLR	Limited to fresh food supply chains
Haji et al. [15]	To review how technologies are implemented across different food supply chain stages and assess their effectiveness.	137	2000–2020	LR	Focus on I4.0 technologies
Rejeb et al. [30]	To review existing digital transformation initiatives in food supply chains, including blockchain, artificial intelligence, big data, social media, and geographic information systems.	2140	1975–2021	BR	Limited to the scope of I4.0 technologies with a specified focus on food supply chains
Yadav et al. [31]	To review the current applications of I4.0 technologies for research into the agricultural food supply chain and to identify current challenges and future research agendas in this area.	146	2010–2020	SLR	Only focused on I4.0 and limited to the scope of the agricultural food supply chain
Remondino and Zanin [32]	To examine the current challenges in the logistics of agri-foods and to present a case study in Italy to demonstrate the importance of digitalization in the logistics of agri-food.	Not provided	Not provided	LR, CS	Limited to the consideration of I4.0 technologies in the supply chains within the agri-food sector
Zhao et al. [33]	To identify and understand the drivers of I4.0 deployment unique to the agriculture food supply chain’s sustainability.	56	Not provided	SLR	Focused on the I4.0 technologies applied in the supply chain of agri-food.
Ertz et al. [34]	To investigate the effects of digital and sustainable technologies on the cold chain sector within the framework of cold chain 4.0.	618	1991–2020	BR, N	Considered the cold chain
Singh et al. [35]	To investigate the relationship between I5.0 and sustainability, with particular emphasis on computational advancements, resource utilization, and environmentally responsible practices in the food industry.	168	All years are covered in the database	SLR	Considering the food industry beyond the supply chain context
Math et al. [36]	To review the published papers on healthcare supply chain management and emerging technologies.	142	2018–2024	BR	Limited to the healthcare supply chain, with consideration of a few technologies such as AI, IoT, and robotics
Tavakkoli-Moghaddam et al. [37]	To examine the role of IoT in the food supply chain and assess its advantages and drawbacks.	93	2014–2021	LR	Scoped to the food supply chain, with consideration of only IoT technology
This study	To provide the most up-to-date review of the state of the art in sustainability-oriented logistics for perishable products within the specific context of I5.0-driven technological transformation.	104	2021–2025	SLR	Novel contributions include: (1) Considering perishable products in a broad sense, rather than being limited to specific product categories. (2) Placing stronger emphasis on a wide range of I5.0 technologies.

Notations: SLR: systematic literature review; LR: literature review; BR: bibliometric review; CS: case study; N: network analysis; WoS: Web of Science.

.

The previous discussions of the existing literature indicate that, at present, reviews concerning perishable logistics, sustainability, and I5.0 are characterized by distinct research silos. A vast majority of studies remain product-centric, relying on specific categories, such as agri-food, fresh foods, and healthcare, while paying limited attention to the most recent technological developments in perishable logistics. Research into sustainability remains largely confined to the framework of I4.0. Although two recent studies [35,36] reveal a growing interest in I5.0, they lack a cross-sectoral perspective concentrating on a particular product category. This gap has motivated the present study, which aims to deliver a contemporary systematic analysis of the sustainability–technology nexus in perishable goods logistics. This paper examines how I5.0 technologies serve as catalysts for sustainable transformation encompassing diverse perishable products. Special attention is devoted to identifying current research trends, emerging themes, and application areas, as well as future research directions for integrating I5.0 into more resilient and sustainable perishable logistics.

1.2. Paper Organization

The rest of this paper is organized into four main parts. Section 2 establishes the theoretical framework, covering sustainable logistics for perishables alongside I4.0 and I5.0. Section 3 defines the core research questions and details the methodology, specifically the systematic literature review and bibliometric analysis. Section 4 presents an in-depth analysis of the collected data to answer the research questions, while the conclusions drawn from the review are summarized in Section 5.

2. Theoretical Background

2.1. Sustainable Perishable Logistics

Logistics is defined as an important subsystem of supply chain management, encompassing the strategic orchestration of raw materials sourcing, inventory management, and the transportation of goods from pre-production to post-production [38,39]. Generally, traditional logistics prioritizes the effectiveness and efficiency of the whole system since they directly influence an enterprise’s performance from the perspective of costs, customer satisfaction, and profitability.

Owing to increasingly severe environmental and social challenges, incorporating sustainability into logistics has attracted much attention, thereby necessitating a broader evaluative framework. Theoretically, sustainable development is generally characterized by three core pillars, economic growth, environmental protection, and social fairness, with the aim of fulfilling contemporary needs without compromising the ability of future generations [40]. This triple bottom line has supported the concept of sustainable logistics transitioning from predominantly focusing on environmental dimensions, such as minimizing ecological impacts [41], to the expanded triple dimensions, involving economic and social metrics, such as employment opportunities and labor standards, and further requiring harmonizing socio-economic outcomes with environmental health [42].

In the specific context of perishable goods, as discussed in Section 1, the inherent trade-offs between economic requirements (e.g., rapid delivery) and environmental burdens (e.g., high-emission refrigeration) necessitate a more sophisticated decision-making process. Consequently, sustainable perishable logistics is no longer a simple question of improving operational efficiency but rather the synergy between network design, distribution, sourcing, and resource allocation to minimize waste and environmental degradation while ensuring social and economic values. To meet such a requirement, novel technologies are fundamental.

2.2. Transition from I4.0 to I5.0

I4.0 was initially introduced in the early 2010s as a strategic initiative aimed at driving technological and industrial transformation [43]. Its fundamental concept centers on “smart manufacturing,” in which a wide range of advanced technologies are integrated to enhance production efficiency and cost-effectiveness through automated, interconnected, and intelligent systems [44,45]. These technologies include connectivity and data-driven tools such as the Internet of Things (IoT), cloud computing, and big data analytics; intelligent processing enabled by artificial intelligence (AI) and cyber-physical systems; as well as physical innovations such as autonomous robots, unmanned aerial vehicles, and additive manufacturing. Additionally, enabling technologies such as blockchain, virtual systems, simulation, and cybersecurity are also leveraged within this paradigm [42].

Despite the efficiency and productivity achieved under I4.0, the rapid shift toward increasingly dehumanized and automated production systems has raised concerns among workers, society, and policymakers [46]. In the context of I4.0, humans (workers) are progressively relegated to peripheral roles, their agency subordinated to the demands of the manufacturing system. Technological monitoring, real-time communication, and simulation are deployed to maintain system-level resilience. This observation reveals a structural limitation of I4.0: resilience is confined to the robustness of technological systems, leaving aside organizations, businesses, and other human-dependent decision-making across the entire framework. It is precisely this limitation that has motivated the emergence of I5.0.

First introduced in 2021 by the European Commission (EC, 2021), I5.0 was defined as a paradigm that “complements the existing I4.0 approach by highlighting research and innovation as drivers for a transition to a sustainable, human-centric, and resilient European industry” (pp. 3–4). This definition has clarified a fundamental reorientation: I4.0 technologies are not replaced but redirected for advanced purposes. As illustrated in Figure 1, core technologies (big data, IoT, blockchain, AI, and collaborative robotics) that emerged in the era of I4.0 illustrate this concept of reorientation most concretely. What distinguishes I5.0 is not solely the introduction of new technologies, but what they are made to serve. These technologies in I4.0 are utilized for system efficiency, cost reduction, and the minimization of human intervention, whereas I5.0 redirects them toward human empowerment, social fairness, and socio-environmental sustainability within production systems [47].

As a nascent paradigm, I5.0 remains subject to academic debate regarding its technical boundaries and distinction from I4.0 [48], with ongoing research exploring how human ideas, real-time data, cybersecurity, and automation can be systematically integrated into existing I4.0 technological frameworks. At the same time, it remains an open question how broader socio-environmental benefits can be distilled from purely economic or productive performance metrics. We provide the following descriptions of five widely adopted technologies [49,50] to trace how the transition from I4.0 to I5.0 manifests within each of them:

• Big data [51,52]: Under I4.0, big data is deployed to extract patterns from large-scale datasets, identifying hidden patterns and correlations within data to generate competitive advantages for organizations [53,54]. Under I5.0, big data is a strong tool leveraging real-world information in enhancing human–machine collaboration, enabling data-driven processes, and supporting the development of personalized products and services.
• IoT: Connecting devices globally is the core value of IoT. In the I4.0 framework, it enhances the intelligence of both physical and virtual entities [55], leveraging seamless data exchange and real-time communication, and further improves operational efficiency [56]. As highlighted by Singh et al. [57], in the era of I5.0, the connectivity is reframed with IoT and sensor networks for industrial sustainability, rather than the general enhancement of efficiency.
• Blockchain: It has evolved significantly with a broad expansion of purpose. Originally adopted in I4.0 as a decentralized, secure ledger system for its advantages of transparency and immutability in transactions [58,59], blockchain is repositioned in I5.0 as a technology to promote social fairness and personal data security (data sovereignty). It can incentivize data sharing among users and stakeholders, enhancing user experience and enabling personalized services across supply chain management, healthcare, and digital identity applications.
• AI: In I4.0, AI focuses on training data-driven models to replicate human-level intelligence with minimal human input [44,60]. In I5.0, Leng et al. [61] identify three transformative modes, collaborative intelligence, self-learning intelligence, and crowd intelligence, which shift AI from model optimization toward multi-dimensional human–machine interaction, enabling adaptive and self-improving systems.
• Collaborative robotics: This technology reflects a direct transition from automation toward human–machine synergy. Under I4.0, autonomous robots reduce human intervention in productive tasks, lowering costs and minimizing human errors [62,63]. In I5.0, collaborative robots (cobots) shift from replacing human roles toward human–machine synergy: equipped with trained intelligence, they jointly support decision-making and identify patterns beyond unaided human perception [64].

In summary, Figure 1 depicts the overall paradigm shift across five technologies, emphasizing the coherent reorientation of their purposes within industrial processes. This reorientation is anchored by three pillars shown in the lower panel of Figure 1. Human-centricity holds that industry should be designed and developed to serve not only stakeholders or machines, but also workers and society. Resilience extends the concept of robustness beyond technological infrastructure to encompass the adaptive capacity of processes and organizations in facing external and unforeseen disruptions. Sustainability defines the overarching direction of industrial development, one in which socio-environmental values are treated as primary design objectives rather than secondary considerations. These three pillars form a logical sequence, responding to “who” (human-centricity), “how” (resilience), and “why” (sustainability) in the promotion of new industries. Within this framework, the present study situates its analysis of perishable logistics, a domain in which all three pillars carry direct operational and ethical relevance.

3. Research Questions and Methodology

Based on the background discussed in Section 1 and Section 2, this review has refined the following four research questions:

• R.Q.1: What characterizes the scholarly landscape regarding perishable goods logistics within the context of I5.0 and sustainability? Specifically, which nations, organizations, researchers, and journals represent the most significant contributors?
• R.Q.2: What are the dominant and recurring research patterns currently shaping this specific field of study?
• R.Q.3: What are the emerging research themes in perishable product logistics toward sustainability and I5.0 integration?
• R.Q.4: What are the future research perspectives on sustainability practices and I5.0 technologies for transforming perishable product logistics?

To address the key research questions, we adopted a systematic literature network analysis approach proposed by Colicchia and Strozzi [65]. This methodology follows a dual-stage process that integrates a comprehensive literature review with a detailed bibliometric network analysis. These are briefly summarized below, and the whole research methodology is illustrated in Figure 2.

3.1. First Phase: Systematic Literature Review

To ensure scientific rigor and maintain transparency, this systematic review follows the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines strictly [66]. The review process followed a structured three-stage protocol: initially identifying relevant literature, subsequently screening the retrieved articles against specific criteria, and finally determining the definitive set of papers for inclusion.

3.1.1. Identification of Literature Related to the Study

(a)

Database selection

The selection procedure is conducted across original research articles published in high-impact, peer-reviewed journals in the fields of logistics, engineering, and sustainability (environments), indexed in two prominent databases: Web of Science (WOS) and Scopus.

(b)

Identification of keywords

To capture the intersection of I5.0 and sustainable perishable logistics, we organized the keywords into four functional groups. The keyword combination followed the “macro domain with detailed subject + target topics” logic. The first group indicates the highest and principal layer of research (i.e., domains), including one keyword: “Logistics”. The second group narrows the research frontier to specific subjects, involving keywords, such as “Perish*” and “Perishable products”. To broaden the search scope, “target topics” are divided into two groups: one focuses on technologies, using keywords such as “Industry 5.0”, “Technology”, “Artificial Intelligence”, “Big Data”, “Robotics”, “IoT”, and “Blockchain”, while the other concentrates on “Sustainab*”.

(c)

Search strategy

The keywords chosen above were used with Boolean operators like “AND” and “OR” to refine the results and specificity. Considering the up-to-date nature of a review, the search was refined to the period from 2021 (i.e., when the topic was first introduced) to 2025, and limited to English articles. The final search strings utilized are shown in

Table 2. Search strings used to identify the relevant studies.

No.	Search Strings
1	“Logistics” AND (“Perishable products” OR “Perish*”) AND (“Sustainab*”)
2	“Logistics” AND (“Perishable products” OR “Perish*”) AND (“Technology”)
3	“Logistics” AND (“Perishable products” OR “Perish*”) AND (“Industry 5.0”)
4	“Logistics” AND (“Perishable products” OR “Perish*”) AND (“Artificial Intelligence” OR “Big Data” OR “Robotics” OR “IoT” OR “Blockchain”)

.

These strings contain different searching purposes, aiming to identify sustainability-related research in perishable logistics, directly capture I5.0 in this realm, and further investigate specific technologies.

3.1.2. Screening of the Articles

(a)

Initial search and refinement

To ensure the integrity of the database and the transparency of the results, the raw data obtained from WOS and Scopus were screened, de-duplicated, and checked to eliminate overlapping records. The initial pool contained 283 articles, which went through a further screening process.

(b)

Inclusion and exclusion criteria

In the screening phase, first, articles’ titles and abstracts were reviewed to assess their conformity with the inclusion and exclusion criteria. In the “eligibility” stage, a more thorough assessment procedure was conducted, in which full texts of the suitable articles were examined for their alignment with this research scope. Specifically, the criteria were applied based on multiple facets, i.e., document types, language, time interval, domains, and research cores. The inclusion and exclusion criteria are the following:

• Inclusion: Peer-reviewed English articles (full-text) published from January 2021 to July 2025, specifically focusing on I5.0 technology or sustainability in perishable product logistics.
• Exclusion: Other types of publication (non-peer-reviewed articles, books, book chapters, conference papers), non-English works, and any research not directly related to the specified domain.

The screening and inclusion process was conducted by the lead researcher, with all decisions subsequently reviewed and validated by the co-authors. As no formal dual-screening protocol was implemented, the absence of a quantitative inter-rater reliability measure, such as Cohen’s kappa, is acknowledged as a limitation of the present study.

3.1.3. Inclusions of Final Papers to Review

Following the application of pre-defined inclusion and exclusion criteria to the initial search results, a total of 104 articles were identified as relevant for comprehensive full-text analysis and data synthesis. A visual representation of this selection workflow is provided in the PRISMA flow diagram (Figure 3). The full PRISMA check list can be found in Supplementary Materials.

3.2. Second Phase: Bibliometric Network Analysis

Based on the resulting relevant set of publications identified from the systematic literature review, a bibliometric analysis is performed to synthesize the large volume of textual data, allowing a rigorous report of descriptive findings on publication trends, key authors and institutions, and research themes. First, the bibliometric analysis was conducted employing Rstudio (Version 2025.05.01+513), using the Bibliometrix package, for performance analysis and science mapping to identify the most-cited countries, authors, and institutions. To perform the country-based co-authorship analysis and keyword co-occurrence analyses, we used the VOS viewer software package. These results are presented in the following section.

4. Discussion

4.1. Answer to R.Q.1

4.1.1. Yearly Publication Progress

Figure 4 illustrates the temporal distribution of research focused on the technological and sustainable advancement of perishable logistics from 2021 to 2025. The data reveal consistent and sustained interest, with an average of approximately 21 publications per year, while 2021, 2023, and 2025 maintained a steady output (23, 24, and 24 articles, respectively). The slight fluctuations observed in 2022 (15 articles) and 2024 (18 articles) do not detract from the overall upward trajectory. This consistent focus reflects a global academic response to the pressures of stringent environmental mandates (such as the 2030 SDGs) and the technological maturation of I5.0 enablers.

4.1.2. Top Productive Authors

The list of the 10 most productive authors in the field, along with their statistics from 2021 to 2025, is shown in Figure 5. It is apparent that, in terms of research output, Tavakkoli-Moghaddam R and Wang X stand out, each having achieved relatively high citation counts and a publication peak in 2023. Following these two researchers, Cimen M., Chen Q., and Manzardo A. have also made notable contributions, particularly in 2025. Early pioneers in this field are Lam Hy and Mangl Sk, who initiated research contributions as early as 2021. Overall, this indicates that only a small group of researchers has been consistently active in the field, with varying degrees of influence as reflected by their citation counts. Most researchers began their active contributions between 2023 and 2025, underscoring the novelty and emerging nature of this research topic.

4.1.3. Most Productive Countries and Institutions

The top ten countries and affiliations are shown in Figure 6 and Figure 7, respectively. As illustrated in Figure 6, China, Iran, and India are the three most productive countries in terms of research output of I5.0-driven and sustainability-enhanced perishable logistics, with 30, 12, and 11 publications, respectively. Notably, these countries are classified as developing nations, yet they demonstrate rapidly growing technological capabilities on the global stage. Among the top ten, only Italy, Belgium, and France are developed countries, contributing eight, three, and two publications, respectively. Figure 7 further shows that the Islamic Azad University (Iran) and the University of Tehran (Iran) are the most productive institutions, each contributing nine papers. They are followed by Hacettepe University (Turkey) with seven papers, and the Hang Seng University of Hong Kong with six papers published. This distribution highlights a concentration of research output from institutions in the Middle East (Iran and Turkey), along with significant contributions from East Asia (China and Hong Kong), consistent with the list of the most productive countries. This distribution may suggest that these regions are more interested in I5.0 and sustainable perishable logistics, likely driven by the urgent need for industrial upgrading to respond to the high demands for perishable products and meet growing environmental requirements.

4.1.4. Citation Patterns of Top Journals

Figure 8 shows the top ten journals publishing research on sustainable perishable logistics. While a total of 69 sources were identified in this field, the concentration of publications in these leading venues reflects the interdisciplinary nature of the field, covering environmental science, engineering, manufacturing, and management. In terms of publication output, Sustainability was the most prominent journal, publishing six articles, followed by the International Journal of Production Research with five articles. Other notable journals include Expert Systems with Applications, Food, and the Journal of Modelling in Management, each contributing four articles. In terms of citations, as can be seen from

Table 3. Top 10 sources that published articles in the field of sustainable and technology-driven perishable product logistics.

Source	NP	h_Index	g_Index	m_Index	TC	PY_Start
Sustainability	6	6	6	1.5	99	2022
Expert Systems with Applications	4	4	4	0.8	94	2021
International Journal of Production Research	5	4	5	0.8	161	2021
Foods	4	3	4	0.6	46	2021
Journal of Modelling in Management	4	3	4	0.6	62	2021
Computers & Industrial Engineering	3	2	3	0.4	37	2021
Environmental Science and Pollution Research	2	2	2	0.5	26	2022
IEEE Access	2	2	2	0.4	97	2021
International Journal of Production Economics	2	2	2	0.4	22	2021
Journal of Cleaner Production	3	2	3	0.4	206	2021

NP: number of publications, TC: total citations, PY_start: publication year. Source: compiled by the authors using the Bibliometrix JR package.

, the Journal of Cleaner Production is the most influential journal, with 206 citations in the area of sustainable perishable logistics. Other highly cited journals include the International Journal of Production Research, Sustainability, and IEEE Access.

Figure 9 presents the results of a three-field analysis using a Sankey diagram, illustrating the dynamic linkages among three key elements in the study: countries (AU_CO), keywords (DE), and journals (SO). In the diagram, the width of each flow band represents the relative contribution of a given element. Journals are displayed on the left, the most frequently used author keywords appear in the center, and the most active contributing countries are shown on the right. For example, as shown in Figure 9, the leading countries in research on sustainability-oriented and technology-driven perishable logistics are China, Iran, and Italy. They all demonstrate a wide focus from “Internet of Things” to “food safety” for this topic. The links among countries, keywords, and journals imply that “Sustainability”, “Supply chain”, and “Optimization” emerge as the most important themes.

4.2. Answer to R.Q.2

This section utilizes three distinct network analysis techniques (i.e., semantic, citation, and collaboration analysis) to map the connections between researchers, thematic clusters, and published documents. Through this, we aim to uncover the prevailing research trajectories for sustainable logistics of perishable products in the I5.0 era.

4.2.1. Semantic Network Analysis: Keywords’ Co-Occurrence

The semantic network analysis is primarily based on the keywords’ co-occurrence. As reported in

Table 4. Top 11 keywords used in the field of sustainability-oriented and technology-driven perishable product logistics.

Keyword	Cluster	Betweenness Centrality	Closeness Centrality	Occurrences	Total Link Strength
Management	Management	1763.6	0.00225	25	304
Optimization	Optimization	1252.3	0.00200	22	281
Logistics	Management	1314.9	0.00228	19	247
Model	Optimization	1999.1	0.00206	20	242
Perishable products	Optimization	2111.2	0.00228	19	221
Sustainability	Optimization	1843.5	0.00210	17	213
Supply chain	Management	2068.0	0.00234	17	174
Quality	Management	1175.3	0.00223	14	144
Algorithm	Optimization	752.1	0.00199	12	137
Design	Management	591.2	0.00213	10	129
Cold chain	Management	1038.4	0.00220	10	125

, “Management” is the most frequently used keyword by the authors in research related to perishable product logistics, with a total of 25 occurrences and a total link strength of 304. “Optimization”, “Logistics”, “Model”, and “Perishable products” are also widely used, with total link strengths of 281, 247, 242, and 221, respectively. In Figure 10, the visual representation of keywords is grouped into various clusters. Notable clusters are red, green, purple, blue, and yellow. The red cluster may indicate a long-standing focus on management, especially inventory control under uncertainty. The green cluster emphasizes the specific requirements of time windows in perishable logistics. The yellow and blue clusters both involve the cold chain. The yellow cluster represents the basic protocol, such as maintaining food quality, while the carbon footprint is first mentioned in the blue cluster, suggesting the importance of establishing more ecological transport for fresh products. The purple cluster exclusively concentrates on e-commerce. A notable finding is the co-occurrence of “China”, which is the country with the most active e-commerce in the world. This suggests that when e-commerce enables long-distance transport of perishable products, the adoption of new technologies to prevent food losses becomes inevitable. It should be noted that “Efficiency” becomes one of the most marginalized keywords in this Figure. This may indicate that for perishable logistics under I5.0, improving efficiency is no longer a core consideration, and social costs based on (product) “quality”, “waste”, “security”, and “system” (resilience) must be addressed. Notably, keywords related to data-driven technologies barely appear. “Machine learning” is set on the edge, while “big data” and “digital twin” are categorized in the yellow cluster. This suggests that the cold chain is the area where accurate data is more urgently needed. The current literature still focuses on the use of “optimization model” under the “management” realm. The use of data-driven methods can be regarded as possible future research and practical directions.

A closer examination of betweenness centrality reveals an important structural distinction among the top keywords. The two highest-ranking terms, perishable products (2111.2) and supply chain (2068.0), are framing concepts of the domain. They are broad enough to be adopted across diverse research contexts, enabling them to serve as bridges connecting multiple research communities. By contrast, management, despite recording the highest occurrence frequency (25), ranks only fifth in betweenness centrality; optimization presents a similar pattern, with the second-highest frequency yet a comparatively low betweenness centrality. This reflects a degree of insularity in both clusters: as optimization is rooted in management science, these methodologically specific terms tend to remain within their research communities and exhibit limited generalization across the broader network.

The clustering coefficients of these two keywords in

Table 5. Comparison of centrality and density of clusters within four quadrants.

Quadrant	Cluster	Clustering Coefficients	Callon Centrality	Callon Density
Motor	Management	0.1566426	17.61	70.10
Motor	Optimization	0.1665161	17.62	58.54
Motor	Demand	0.3707665	5.72	91.59
Motor	IoT	0.6282051	4.43	88.58
Niche	Carbon footprint	0.8105263	0.25	62.50
Niche	Emission reduction	0.6000000	0.25	62.50
Basic	Models	0.3247863	1.07	33.33
Emerging	Big data	1.0000000	0.00	50.00
Emerging	Machine learning	-	0.50	50.00
Emerging	Sustainable development	0.4967742	0.00	50.00

have further corroborated this insularity. Optimization (0.167) and management (0.157) exhibit the lowest clustering coefficients among all substantive keywords, confirming that despite their high frequency, their neighbors are weakly interconnected. Specifically, the keywords co-occurring with optimization are almost exclusively drawn from operations research—algorithm, model, inventory, time windows, and vehicle routing—forming a tightly bounded silo that reflects the dominance of using a single quantitative method rather than cross-disciplinary engagement. Management, by contrast, carries greater potential for integrating the human-centric and resilience pillars of I5.0, which require contributions from organizational studies, business analysis, and policy research to provide qualitative definitions. Yet its low clustering coefficient suggests this potential remains unrealized, as the concept is currently deployed primarily in its operational sense. This structural gap points to an important future direction: advancing I5.0 in perishable logistics will require moving beyond mathematical modeling toward management research that investigates policy design and questions of social sciences.

A counterexample is sustainability. It records a lower occurrence frequency (17) and total link strength (213) than management (25 and 304, respectively) yet achieves a higher betweenness centrality (1843.5 vs. 1763.6) and belongs to the optimization cluster rather than the management cluster. This indicates that sustainability functions not only as a high-frequency research topic, but also as a bridge between different research clusters. As the earliest of the three I5.0 pillars to gain attention in the literature, sustainability has evidently diffused across the full breadth of the field.

Finally, the closeness centrality values of all keywords fall within a narrow range (0.00199–0.00234), suggesting that despite their thematic diversity, the core concepts of this field remain mutually close within the network. All the metrics indicate that research attention is concentrated around a relatively small and interconnected set of focal themes.

4.2.2. Citation Network Analysis: Research Themes

Figure 11 shows the outcomes of the thematic map analysis, examining the conceptual and functional framework of the reviewed studies to identify key research themes, textual insights, and prominent clusters within the field. By analyzing author keywords, the strategic diagram categorizes themes into four types according to their centrality and density: motor themes (upper right), niche themes (upper left), basic themes (lower right), and emerging or declining themes (lower left) [67]. In this context, centrality indicates the level of interaction or connectivity of a theme within the field (i.e., its relevance across studies), while density reflects the internal development of the theme (i.e., its coherence and maturity). Table 5 reports the centrality, density and clustering coefficients underlying the thematic map for all identified clusters, providing a quantitative basis for the following discussion.

In the upper-right quadrant of the strategic diagram, the terms “perishable goods”, “internet of things”, “management”, “logistics”, and “supply chain” can be identified as the most popular keywords in the motor theme, as they are highly recurrent, well-developed, and highly relevant to the research area. Notably, the demand cluster records the highest Callon density among all clusters (91.59), indicating an exceptionally cohesive research topic organized around inventory problems, while the IoT cluster (88.58) reflects a similarly mature and internally focused IoT research stream. In contrast, the terms “machine learning”, “fresh”, “efficiency”, “big data”, “sustainable development”, “temperature” and “shelf life” appear in the lower-left quadrant, classified as emerging or declining themes. These clusters exhibit both low centrality and density values at or below 50, confirming that they remain weakly connected to the broader network and insufficiently developed internally. Of particular significance is the position of ‘big data’ (centrality: 0, density: 50) and ‘machine learning’ (centrality: 0.50, density: 50), which together represent the data-driven technological enablers most closely associated with the I5.0 vision, yet are structurally the most marginal within the current research landscape. Moreover, “China”, “electronic commerce”, “sustainable agriculture”, “truck vehicle”, “carbon emission” and “production strategy” appear in the upper-left quadrant as niche themes. The “carbon footprint” and “emission reduction” terms are located in this quadrant as well. Both clusters share a density of 62.5 but near-zero centrality, suggesting that environmental impact research has developed a self-contained discourse that has not yet been structurally integrated into the management and optimization streams. Lastly, the lower-right quadrant includes “cold storage cost”, “models”, “food safety” and “perishable systems”, which are identified as basic themes. The particularly low density of the “models” cluster (33.33, the lowest among all clusters) indicates that methodological approaches remain highly fragmented.

Regarding the clustering coefficients, at the global level, 0.411 indicates moderate overall cohesion. As evidenced by their low clustering coefficients (management: 0.157, optimization: 0.167), the two dominant motor clusters remain loosely connected internally despite their high frequency. By contrast, technologically oriented themes such as IoT (0.628) display significantly higher clustering, reflecting the emergence of well-defined and internally consistent subfields. Niche themes, including carbon footprint (0.811) and emission reduction (0.600), demonstrate strong internal cohesion but low centrality, indicating that environmental sustainability research is well-developed yet remains structurally peripheral to dominant management and optimization streams. Finally, big data exhibits a clustering coefficient of 1.0, which paradoxically signals not maturity but isolation: this term co-occurs only within an exceedingly small, tightly closed group of keywords, characteristic of an early-stage research area that has not yet achieved broader network integration.

The absence of machine learning from the clustering coefficient analysis reflects a broader terminological shift. As a methodological term, machine learning stands for a specific subset of data-driven approaches and has increasingly been superseded in academic discourse by broader terms such as data-driven and artificial intelligence. Studies employing machine learning methods may therefore appear in the network under alternative keyword labels, suggesting that the actual penetration of predictive and learning-based methods in this field may be greater than the keyword frequency data alone would indicate.

4.2.3. Citation Network Analysis: Trending Topics

Figure 12 presents the trending topics in sustainable perishable logistics research from 2021 to 2025, with a focus on technological advancements. The analysis indicates that early research during this period primarily addressed topics such as “blockchain,” “Internet of Things,” “time,” “design,” “systems,” and “management.” By 2022, the literature began to incorporate discussions on “quality,” “logistics,” “model,” “impact,” “algorithms,” and “optimization,” with optimization remaining the most prominent research topic across the entire period. Notably, “cold chain logistics” appeared as the most recent research focus in 2025 within this domain.

4.2.4. Citation and Co-Citation Analysis

This subsection employs citation and co-citation analysis to explore the networks connecting authors, countries, and institutions engaged in research on perishable product logistics. In particular, citation analysis assesses the relative importance or influence of a research article, author, journal, or country by measuring how frequently they are cited [68]. Co-citation analysis, on the other hand, examines instances where two documents are cited together by other documents [69], measuring the combined influence of the two articles. Specifically, the greater the number of co-citations between two publications, the more closely related they are [70].

Figure 13 presents the citation analysis network of 104 articles, revealing that the papers by Jouzdani and Govindan [71], Tsang et al. [72] and Wang et al. [73] are among the most highly cited. Figure 14 visualizes the co-citation map, where nodes represent articles, and links indicate co-citation occurrences. The distance between nodes reflects the degree of topical relatedness [74]. As shown in the map, seven distinct clusters, colored red, green, blue, yellow, purple, cyan, and orange, represent groups of articles that are often cited together, suggesting shared research themes based on subject-relatedness or keyword similarity.

4.3. Answer to R.Q.3

To examine the structural patterns of the reviewed literature and identify emerging research themes, a cluster analysis based on bibliographic coupling was conducted. The resulting clusters are shown in Figure 15 while the leading articles of these clusters are shown in

Table 6. Author clusters based on the reviewed articles.

Cluster Articles	Link	Total Link Strength
Cluster 1: Perishable product quality and safety
Abbas et al. [75]	37	79
Fan et al. [76]	45	78
Gillespie et al. [77]	36	75
Skawińska and Zalewski [78]	27	59
Defraeye et al. [79]	22	45
Cluster 2: Sustainability-oriented management
Shaharudin and Fernando [80]	32	47
Pilati et al. [81]	31	45
Dhanda et al. [82]	16	23
Zhu et al. [83]	19	23
Kumar et al. [84]	13	13
Cluster 3: Resilient supply chains and logistics under uncertainty
Fasihi et al. [85]	44	102
Jouzdani and Govindan [71]	40	86
Heidari et al. [86]	37	66
Pan et al. [87]	36	65
Navazi et al. [88]	32	64
Cluster 4: Transformation toward I5.0-assisted monitoring
Chen et al. [89]	38	83
Kumar et al. [90]	36	66
Zou et al. [91]	31	56
Jarumaneeroj et al. [92]	30	50
Lam et al. [93]	27	41
Cluster 5: Sustainability-targeted optimization frameworks
Leng et al. [94]	34	100
Leng et al. [95]	25	85
Golestani et al. [96]	39	71
Wang et al. [73]	29	57
Abbasi et al. [97]	23	47
Cluster 6: Sustainable routing problems for perishable product logistics
Jahdi et al. [98]	44	92
Köseli et al. [99]	24	52
Majidi et al. [100]	23	51
Cluster 7: Technological and security barriers
Bhutta and Ahmad [101]	5	7
Sabbagh [102]	1	1

Source: Compiled by the authors using VOS Viewer.

. For each cluster, representative articles are selected based on their total link strength (TLS), which reflects the degree of bibliographic coupling within the cluster and is therefore a suitable selection criterion. A complete list of the 104 included studies is provided in Appendix A. Their relevant description and discussion are presented in the following sections.

4.3.1. Cluster 1: Perishable Product Quality and Safety

Food quality and safety, as non-separable components of a sustainable food system, significantly influence not only economic outcomes but also societal and environmental dimensions. The studies in this cluster (cluster color: red) trace the technological evolution of perishable logistics. Within the investigated time interval (2021–2025), the early work, such as Fan et al. [76] (TLS of 78), concentrated on using numerical methods (agent-based simulation) to optimize the trade-offs between cost, emissions, and quality in cold chain design. Notably, the concept of “product and management quality” addressed in this work remained within the I4.0 paradigm, which is largely confined to the economic performance of the supply chain. Published in the same year, Defraeye et al. [79] (TLS of 45) examined the key advantages of digital twins, mostly based on AI, for fresh horticultural produce supply chains. The study found that digital twins are highly effective in identifying and forecasting supply chain issues that affect food quality and cause food loss. An interesting proposal that is related to I5.0 is using the analogous framework based on customized digital twins in healthcare and pharmaceutical logistics, where the shelf life of products would serve as a principal indicator for intelligent shipping. These two works represent a technological prototype aligned with I5.0 principles, yet they remain theoretical, lacking robust qualitative or quantitative validation.

Skawińska and Zalewski [78], with a total TLS of 59, mark a meaningful point in this trajectory. By accounting for the systematic disruptions, such as policy barriers, instability or delays in transportation, introduced by the COVID-19 pandemic, the authors identified the need for data-based real-time temperature monitoring to address dynamic changes in perishable product logistics, resulting in a framework that integrates the Statistical Process Control (SPC) charts into IoT and RFID. This work signals the initial utilization of historical data to support real-time perception of logistics practitioners.

In the subsequent studies published in 2023, the leading article by Abbas et al. [75] (79 TLS) shifted attention towards the on-field cold chain, which requires more intensive real-time control to mitigate the ‘field heating’. The authors explicitly considered social impact as an important model outcome assessed through product quality and safety (safe shelf) within a multi-objective optimization. Gillespie et al. [77] (TLS: 75), adopting an even more practical application, deployed an LTE-M-integrated IoT system, specifically designed to reduce errors at the human–technology interface, which reduced false alerts received by warehouse and delivery staff.

Taken together, the studies in this cluster trace a meaningful technological trajectory, collectively establishing the early-stage foundation for I5.0 adoption in perishable quality control. Concerning the triple bottom line, the economic dimension (cost efficiency, waste reduction) and environmental dimension (energy consumption, product loss) are consistently addressed, while the social dimension remains peripheral: safety and quality are framed primarily as operational outcomes rather than as markers of consumer welfare or worker wellbeing. With respect to I5.0 alignment, the sustainability pillar is partially realized through waste reduction and energy efficiency goals, but human-centricity remains constrained: technologies predominantly operate in a unidirectional “data-to-human” mode [77], delivering outputs to operators without incorporating worker expertise into system design. Resilience is acknowledged through COVID-19-motivated disruptions but is not systematically addressed. The key unresolved challenge for this cluster is the design of bidirectional human–machine interfaces, a prerequisite for I5.0 transition that no study in this cluster has yet fully realized.

4.3.2. Cluster 2: Sustainability-Oriented Management

Cluster 2 (green) represents an emerging research theme in the field of logistics and supply chain management for perishable products, emphasizing the enhancement of sustainability across the entire process. As illustrated in Figure 14, the relatively sparse inter-paper connections within this cluster reflect the considerable diversity in how researchers interpret the concept of sustainability, making the cluster heterogeneous.

Like [78] in Cluster 1, Kumar et al. [84] identified the impact of COVID-19 on perishable product supply chains. They analyzed risk-mitigation strategies during the pandemic using the fuzzy best–worst method (F-BWM). The findings indicate that, of the mitigation strategies examined, collaborative management, proactive business continuity planning, and financial sustainability demonstrate the greatest effectiveness. Notably, the study’s emphasis lies on the organizational level rather than the technological one. The idea of maintaining socio-economic sustainability through collaborative management facing exogenous disruptions, such as pandemics, closely aligns with the core I5.0 principles of human-centered collaboration and systemic resilience. However, the absence of any advanced technology to realize these strategies shows a significant gap between this work and the I5.0 paradigm.

More recent studies deal with sustainability from different perspectives. Shaharudin and Fernando [80], a key study in the cluster, explored determinants of cold supply chain management for leafy green vegetables, demonstrating that advanced logistics technologies and coordinated systems (refrigeration, cold storage, and temperature control), significantly enhance freshness, energy efficiency, and overall logistics performance. The study further emphasizes that cold chain logistics, compared to conventional logistics, can benefit more from advanced technologies. Dhanda et al. [82] shift attention to the external policy environment, examining how carbon emission regulations (carbon cap-and-trade policy, carbon buying and selling mechanisms) shape the competitive interactions between suppliers and retailers. This work signals that sustainability discourse should extend beyond operational efficiency to consider external market policies.

Among the more technically oriented contributions, I5.0 remains comparatively limited. Pilati et al. [81] (45 TLS) formulated a bi-objective optimization model with economic cost and greenhouse gas emissions for inventory control. They explored strategies for strengthening sustainability and transparency within decision processes. The study suggests that integrating diverse sustainability considerations into inventory management supports more resilient operations, regulatory compliance, and alignment with increasing demand for green products. Zhu et al. [83] introduced a blockchain-enabled smart contract to tackle the facility location–routing optimization problem, promoting resource sharing and boosting the efficiency and sustainability of fresh product supply chains. Although [81,83] both considered transparency, they still share a common limitation: technology is primarily treated as an instrument to enhance efficiency (I4.0 target), leaving the human–machine collaboration unaddressed. Even though [83] leverages blockchain, the proposed framework does not fully exploit its potential for information transparency and data traceability, which can be regarded as requirements of the I5.0 paradigm of technological interpretability.

The studies in this cluster advance sustainability management through four parallel approaches: organizational resilience [84], technology-driven system design [80], policy coordination [82], and mathematical optimization [81,83]. While this diversity reflects the multi-dimensional nature of sustainability, it also reflects the diversity in how researchers interpret sustainability. Economic and environmental pillars of sustainability are consistently addressed through carbon emission constraints and cost optimization, while the social dimension appears only partially in [84] through workforce continuity and collaborative management. Human-centricity and resilience remain largely unrealized in the sense of I5.0: technology is deployed as an efficiency-oriented instrument rather than as an enabler of human empowerment, and resilience is treated at the organizational level without the technological enablers that I5.0 envisions. Future research may find synergies with the AI and digital twins identified in Cluster 1, potentially bridging the gap between the conceptual breadth of sustainability and the technological infrastructure needed to realize it under an I5.0 framework.

4.3.3. Cluster 3: Resilient Supply Chains and Logistics Under Uncertainty

The research problems addressed in this cluster concentrate on the resilience and responsiveness of logistics and supply chain networks under uncertainty (e.g., demand variability, product perishability, and transit times). Therefore, this research stream focuses on mathematical programming and optimization models as tools for balancing simultaneously the economic, environmental, and social dimensions of sustainability.

The leading article in this cluster (cluster color: blue), by Fasihi et al. [85] (102 TLS), highlighted the increasing interest in sustainability within fish supply chains (SC) and reverse logistics (RL) under uncertain demands. A multi-objective MILP model improves market distribution, reducing costs, enhancing customer fulfillment rate, addressing social concerns, and limiting negative environmental effects such as CO₂ emissions and product waste. Jouzdani and Govindan [71] (86 TLS) introduced a network design framework for food supply chains aimed at minimizing costs, energy use, and traffic congestion under conditions of food perishability and uncertain product lifetime. Heidari et al. [86] (66 TLS) examined the role of forward and reverse food supply chain networks within a socially sustainable supply chain framework. The authors explicitly incorporated economic (cost efficiency), environmental (reduction in greenhouse gas emissions), and social (job creation, community development, and customer satisfaction) dimensions in an optimization model. A similar but more advanced focus is found in the study of Navazi et al. [88] (with 64 TLS), which addresses the sustainable closed-loop location–routing–inventory problem for perishable products. Multi-compartment vehicles are introduced to enable pickup and delivery of forward and reverse flows, retailer and employee satisfaction are integrated into the social objective function, and fuzzy chance constraints are employed to handle demand uncertainty.

Pan et al. [87] (with 65 TLS) represents the most recent advance in technological integration by using IoT to enhance perishable product supply chain networks and logistics. The study proposed an optimization-based framework for a sustainable multi-period perishable supply chain. Service level is defined as a composite of product and on-time delivery rate, bringing the measurement of social sustainability closer to end-user experience. While traditional optimization focuses on costs or simply the minimization of emissions, the integration of IoT-enabled real-time monitoring and mixed fleet configurations into the framework involving social sustainability reflects the initial transition from I4.0 to I5.0.

The studies in this cluster collectively contribute to interpreting and quantifying abstract human-centric parameters, such as satisfaction, employment, and fatigue, which are important I5.0 principles within mathematical programming models. The notable unresolved debate of this cluster is how to integrate core I5.0 technologies into modeling, including digital twins, AI-driven predictive analytics for uncertain disruptions and decision support, and interactive platforms (mobile applications and interfaces) where users can actively inform and refine service providers’ decisions. Currently, these human-centric considerations are incorporated only as discrete parameters within optimization models, rather than as part of a holistic human–machine system. Human-centricity is partially addressed through quantified welfare parameters but remains static rather than dynamic; resilience is engaged reactively through uncertainty modeling without real-time adaptive mechanisms. Regarding the sustainability pillars, the economic and environmental dimensions are well-represented through cost efficiency, network design, CO₂ emissions, and product waste, while the social dimension appears primarily through satisfaction and employment metrics.

4.3.4. Cluster 4: Transformation Toward I5.0-Assisted Monitoring

Within this cluster (yellow), the studies reflect an accelerating transition toward intelligent, real-time monitoring systems for perishable product logistics through AI, digital twins, and IoT. This cluster exhibits a stronger technological kinship with Cluster 1 since an accurate monitor can coherently guarantee the safety and quality of perishable logistics. Works in Cluster 4 can be read as specific applications based on Cluster 1’s considerations, bridging sustainability concerns with intelligent monitoring.

The basic work that established a strategic baseline in this cluster is provided by Kumar et al. [90] (66 TLS), which mapped the critical performance criteria and sub-criteria for a cold supply chain (CSC) using an AHP–Fuzzy TOPSIS hybrid approach. The study identified energy consumption as an important critical performance barrier, followed by environmental impact and product quality, and recommended passive cold devices (PCDs) and IoT integration as priority improvement strategies.

Chen et al. [89] (83 TLS) and Jarumaneeroj et al. [91] (50 TLS) both addressed the above-mentioned challenges at the operational level. Chen et al. [89] advocated for more sustainable cold chain practices for climacteric fruits. The authors introduced a hybrid, retailer-demand-driven multi-objective time–temperature management approach that accounts for trade-offs between retailers’ satisfaction and energy usage alongside post-harvest ripening traits. Jarumaneeroj et al. [91] utilized multi-compartment refrigerated vehicles (MCVs) to enhance eco-friendly long-haul transportation of perishable products. The authors modeled carbon emissions from both driving activity and compartmental cooling thermodynamics, integrating the weight losses of products into the transit time. Lam et al. [93] (41 TLS) proposed a reversible cold chain vehicle routing model that integrates IoT within a multi-objective optimization framework to enable real-time monitoring. These studies show a similar advantage of using optimization models as discussed in Cluster 3: satisfaction, energy consumption or other I5.0 principles can be treated as a component of objective functions or constraints. A clear limitation is that these studies adopt deterministic modeling approaches that limit real-world applicability, and only I4.0 technology (IoT) is considered.

These limitations are partially addressed by Zou et al. [92] (56 TLS). A “five-dimensional model” for cold supply chains was established by merging digital twin (DT) frameworks with AI. Their methodology utilizes a two-stage process to refine temperature forecasting for more precise shelf-life predictions. A notable contribution of this work is the implementation of a Long Short-Term Memory (LSTM) model, which relies on validated experimental temperature data to predict in-box conditions, ultimately reducing product degradation and waste in perishable supply chains. Their proposed framework reflects a foundation for approaching I5.0: using AI-driven techniques to enhance systematic resilience (improving the prediction accuracy), where AI is not employed to approximate economic or efficiency-related variables, but serves to sense and predict the physical state of the product itself. This distinction matters for alignment to I5.0, as it shifts the role of technology from efficiency instrument to quality guardian, contributing directly to the environmental and resilience dimensions of sustainability.

Collectively, this cluster advances the technological foundation of I5.0-assisted monitoring covering multi-criteria analysis [90] and AI [92]. With respect to the I5.0 framework, sustainability is well-addressed across the environmental pillar through energy optimization and carbon modeling, and the economic dimension through cost and quality trade-offs. However, social sustainability remains limited. Human-centricity and resilience only partially emerge, as evidenced by the AI-driven predictive tools in [92]. The cluster’s most significant unresolved debate remains consistent with that identified in Cluster 1: the bidirectional human–machine collaboration is absent. Establishing a data ecosystem spanning consumer preferences, retailer feedback, and real-time demand signals should be a pressing future direction for advancing this cluster to align with fully realized I5.0. Bridging this gap would require not only technical integration across Point-of-Sale terminals, retailer systems, and upstream suppliers, but also a fundamental redesign of how human judgment enters and shapes decision-making.

4.3.5. Cluster 5: Sustainability-Targeted Optimization Frameworks

In this research theme (highlighted in purple), significant efforts have been devoted to balancing cost efficiency, product quality, and environmental impacts within the logistics and supply chain networks of perishable products. In terms of methodology, this cluster closely resembles Cluster 3, as both rely primarily on mathematical modeling and optimization. However, this cluster places emphasis on sustainability while Cluster 3 incorporates more social-level variables.

In terms of research trends, this cluster exhibits a clear trajectory of moving from treating environmental impact as a single cost component towards enhancing systemic sustainability as the principal research emphasis. In earlier studies, specifically [96] and [73], carbon emissions are incorporated into the objective function primarily as a cost term, meaning that sustainability at this stage functions more as an indicator of economic performance rather than a core topic to guide model design. From the perspective of decision-making levels, Golestani et al. [96] proposed a facility location-based MILP at the strategic level while Wang et al. [73] switched to an operational-level multi-center vehicle sharing routing problem. Meanwhile, [73] introduced the concept of fairness using cost-saving methods to consolidate the alliance among cold chain distribution centers. However, within this fairness, the component that should be truly considered, customers, is absent.

As in [78,84] from Clusters 1 and 2, COVID-19 is regarded as a valuable external disruption in [97] (47 TLS). This study represents one of the earliest attempts within this cluster to engage with resilience.

Leng et al. [94] is the leading article in this cluster. The study aims to improve the efficiency and sustainability of urban cold chain logistics by applying a heuristic method to balance economic costs, fuel consumption, product quality loss, and customer satisfaction. The design of six distinct customer satisfaction evaluation functions gives the proposed method a degree of human-centric character. Based on this, Leng et al. [95] (85 TLS) introduces a hybrid electric vehicle fleet alongside a Q-learning hyper-heuristic algorithm, reflecting an initial application of adaptive intelligence at the algorithmic level. Nevertheless, Q-learning serves solely as a selectable evolutionary operator rather than a technique for supporting dynamic decisions or expert-guided reinforcement learning, the latter of which would represent genuine human expertise–algorithm coupling.

This cluster contributes a clear developmental trajectory in sustainability-targeted optimization, progressing from pure cost-efficient models [96] toward frameworks that partially incorporate social fairness considerations [73,94]. With respect to the I5.0 fulfillment, sustainability is well-represented in this cluster, reflecting its growing conceptual maturity in transportation: economic, environmental, and social objectives are integrated simultaneously but imbalanced. Economic and environmental pillars are classical measures in the research of logistics and are well-addressed. The social dimension still appears through customer satisfaction and fairness considerations [73,94], without the discussions of workforce and community welfare. Notably, this cluster retains the unresolved tensions shared by other optimization-oriented clusters: the effective incorporation of dynamic decision-making, such as digital twins and AI-driven approaches, remains largely unrealized. Considering other components of the I5.0 framework, human factors are still incorporated as static model parameters, and resilience is engaged only reactively through COVID-19 disruption scenarios [97].

4.3.6. Cluster 6: Sustainable Routing Problems for Perishable Product Logistics

Within this cluster (in cyan), an emerging research theme focuses on enhancing sustainability within the routing decisions of perishable product logistics. Jahdi et al. [98] (92 TLS) formulated a mixed-integer programming (MIP) model to tackle a multi-period inventory routing problem (IRP) characterized by fluctuating stochastic demand, costs tied to specific routes, and environmental considerations. Specifically, the study sought to harmonize replenishment schedules and delivery paths to achieve a balance between economic cost reduction and the minimization of CO₂ emissions. Köseli et al. [99] investigated a stochastic closed-loop inventory routing problem (IRP) that integrates both ecological and social dimensions into daily food logistics. Their work focuses on reducing CO₂ emissions from refrigerated transportation, managing waste collection and disposal, and improving employee work schedules. Employees are permitted to adjust start times, as the model constrains the longest delivery route to compress the overall time window. Employee rescheduling, regarded as social welfare, is first considered among all studied clusters. On the other hand, Majidi et al. [100] (with a total TLS of 51) tackled a multi-objective optimization problem for perishable goods, integrating sustainable pricing, manufacturing, staffing, and distribution. By concurrently evaluating overall profitability, labor management, and fleet fuel efficiency, the research sought to align operational decisions with broader sustainability targets.

This cluster makes a notable contribution by being the first among all clusters to formally incorporate workforce welfare through employee scheduling and labor management as a social sustainability objective [99,100]. The other two sustainability dimensions are represented through cost and profitability at the economic level, and CO₂ emissions and waste at the environmental level. With respect to I5.0 alignment, the introduction of workforce welfare offers a new facet of human-centricity; however, the absence of resilience represents a notable gap. The unresolved tension is that routing problems represent a natural experimental field for IoT-driven real-time demand sensing, digital twins, and AI-assisted adaptive routing, yet none of these technological enablers are engaged in this cluster. The practices in this cluster could connect with the monitoring frameworks identified in Cluster 4, but this integration currently remains unrealized. Future research should extend workforce welfare modeling beyond scheduling problems to other optimization contexts and integrate responsive and adaptive mechanisms in routing problems.

4.3.7. Cluster 7: Technological and Security Barriers

Supply chain and logistics systems for perishable goods can greatly benefit from key I5.0 technologies. However, the application of these technologies in perishable logistics and supply chains is still at an early stage of implementation. This cluster (in orange) addresses the pressing challenges involved in overcoming the barriers to their adoption. Notably, a reliable and traceable supply chain management framework that ensures transactional integrity, transparency, and data immutability during the logistics of perishable agricultural products was proposed by Bhutta and Ahmad [101] (7 TLS). This framework integrates IoT and blockchain technologies to enable secure monitoring and reporting, providing a viable alternative to traditional supply chain systems that are often costly in both economic and computational terms. Sabbagh [102] (with 1 TLS) further explored key challenges hindering the implementation of blockchain in perishable goods supply chains and logistics. Using an uncertain modeling approach, the authors highlighted technological and security constraints as the dominant barriers. Seven related challenges were pointed out, including the lack of technical maturity, information security concerns, usability issues, unpredictability, lack of cooperation, forking execution, and scalability limitations of blockchain technology.

This cluster contributes a critical examination of the barriers to I5.0 technology adoption in perishable logistics, representing the only cluster explicitly focused on implementation challenges rather than optimization or monitoring. With respect to I5.0 alignment, economic sustainability is partially addressed through cost comparisons with traditional systems, while social and environmental dimensions are neglected. Human-centricity is not directly engaged. The cluster’s primary contribution lies close to resilience, identifying the technological and security preconditions that must be met before I5.0 can be meaningfully realized in practice. The unresolved tension is that the identified barriers, including scalability, security, and interoperability, remain largely theoretical, with no empirical validation of solutions proposed. Future research should move beyond barrier identification toward designing and testing concrete governance and technical frameworks that enable secure, scalable I5.0 implementation in perishable logistics.

4.4. Answer to R.Q.4

Based on the examination of current research themes and the preceding discussion, the widespread adoption of I5.0, particularly in the field of perishable product logistics and supply chains, continues to face several challenges despite its strong potential. High investment costs associated with infrastructure upgrades, compatibility issues with legacy systems, and concerns related to data security and privacy remain significant barriers. Furthermore, workforce upskilling, interoperability across technologies, and the lack of standardized frameworks further complicate implementation. Ethical considerations surrounding AI, automation, and their societal implications also require careful attention. Overcoming these challenges requires collaborative efforts among industry, governments, and technology developers.

4.4.1. Critical Appraisal of I5.0 Alignment

A cross-cluster examination of the reviewed literature reveals that the three defining pillars of I5.0, sustainability, human-centricity, and resilience, have been adopted to uneven degrees. Sustainability is the most structurally embedded pillar across all seven clusters in various forms, ranging from carbon emission constraints in optimization models (Cluster 5, 6) to policy-driven frameworks in management research (Cluster 2). However, its concentration remains primarily economic and environmental in orientation, with the social dimension receiving comparatively limited attention across clusters.

Human-centricity, by contrast, remains the least realized pillar in the reviewed literature. Across Clusters 1 through 7, humans are most represented as a static parameter within mathematical models, such as customer satisfaction scores, employee scheduling constraints, or service level, rather than as a dynamic, bidirectional force that shapes and is shaped by technological systems. The unidirectional “data-to-human” support model identified in Clusters 1 and 4 is symptomatic of a broader pattern: technologies are deployed to deliver outputs to human operators, but the reverse pathway, whereby human expertise actively guides system design and algorithmic learning, remains absent. This suggests that many studies nominally classified under I5.0 continue to operate within an I4.0 techno-centric logic, without restructuring the role of humans in their frameworks.

Resilience follows a similar development. In Clusters 2, 3, and 5, uncertainty and disruption are widely considered as challenges. However, most contributions treat resilience as a reactive and static response to external disruptions of the whole system. The capacity of supply chains to reorganize, reform, and respond to unforeseen disruptions in real time is rarely addressed. The fragmented application of digital twins, AI-driven adaptive routing, and real-time decision support across most clusters further confirms that technological resilience, let alone organizational resilience, has not yet been meaningfully integrated into the research agenda.

Taken together, these observations suggest that the I5.0 paradigm is aspirational and at an early stage of realization. This is not to suggest that the reviewed literature represents a rebranding of I4.0 applications. Meaningful progress is evident: the integration of IoT-enabled real-time monitoring, the explicit incorporation of social sustainability objectives into optimization frameworks, and the emerging application of digital twins and AI-driven predictive tools all represent a transition toward I5.0.

The structural marginality of human-centricity and resilience may be understood through a developmental lens. The trajectory of sustainability itself is instructive: it required decades of conceptual refinement and methodological innovation before it could be quantified through indicators such as carbon emission metrics, and SDG measurement systems [40,41]. The bibliometric evidence presented in this study suggests that human-centricity and resilience are currently at an analogous early stage, where they are widely invoked as principles, but the methodological infrastructure needed to translate them into measurable research constructs remains absent. The insularity of the management and optimization clusters further reinforces this gap: operations research tools are well-suited to quantify objectives such as cost minimization and emission reduction but struggle to deal with concepts that resist easy modeling: worker happiness, organizational adaptivity, social equity, and related socio-technical concepts. Closing this gap will require qualitative approaches, including case studies, mixed-methods designs, and policy analyses, that can give empirical context to what I5.0’s human-centric and resilience pillars require in practice. They ensure that technological development remains aligned with ethical and social boundaries.

4.4.2. Research Directions

A cross-cluster synthesis reveals that the path toward a fully realized I5.0 requires systemic consideration across three facets. First, there is a need for technological interoperability between sensing (Clusters 1 and 4) and secure data architecture (Cluster 7). Second, the current trend of treating human factors as static mathematical constraints in optimization models (Clusters 3, 5, and 6) must shift toward dynamic, bidirectional human–machine collaboration. Third, research should focus on active resilience, moving from planning for known risks to real-time adaptation against unforeseen disruptions (Clusters 2 and 3).

To advance the realization of I5.0,

Table 7. Future research directions.

Cluster	Future Research Directions	Category
1	Real-time sensing and intelligent quality control frameworks for perishable products	Practical implementation
	I5.0 technologies for perishable food safety management and supply chains	Methodological innovation
2	Emerging technologies for sustainable perishable supply chains and logistics	Methodological innovation
	Readiness of perishable logistics and supply chain against pandemics and disruptions	Theoretical development
3	Exploration of technology-supported systems for operationalizing sustainability and resilience in real-world perishable logistics	Theoretical development
	Resilient and flexible network design under uncertainty and potential disruptions	Practical implementation
4	AI-enabled digital twins to enhance predictive quality control and sustainability	Methodological innovation
	Human–machine collaboration in the operation of perishable supply chains and logistics	Theoretical development
	Integration of I5.0 technologies for intelligent and autonomous perishable product supply chains and logistics	Methodological innovation
5	Trade-offs between economic, environmental, and social objectives in perishable supply chain and logistics design	Theoretical development
	Holistic design frameworks with economic, environmental, and social sustainability constraints	Practical implementation
6	AI-driven routing, IoT-enabled monitoring, and decision support for energy-efficient transportation	Methodological innovation
	Green vehicle routing problems for perishable products	Theoretical development
7	Solutions to overcome security barriers to implementing I5.0 technologies in perishable product logistics	Methodological innovation
	Smart human–machine systems for secure, traceable, and efficient perishable supply chains	Practical implementation
	Enhancing information and financial security in digitalized logistics and supply chain environments	Theoretical development

summarizes the key research topics proposed for future investigation across each cluster. These research directions are classified into theoretical development, methodological innovation, and practical implementation.

Given the widespread importance of perishable products and the critical role of their associated logistics and supply chains in supporting human life, the research topics outlined above warrant prioritized research attention. In light of the current volatile global environment, particularly the disruptions caused by regional conflicts that significantly affect logistics activities worldwide, greater priority should be given to the following areas: “Readiness of perishable logistics and supply chain against pandemics and disruptions” (Cluster 2); “Resilient and flexible network design under uncertainty and potential disruptions” (Cluster 3); and “Human–machine collaboration in the operation of perishable supply chains and logistics” (Cluster 4). Accelerating the application of I5.0 technologies in these areas will be critical to fostering more stable and resilient supply chains for essential perishable products.

5. Conclusions

The pursuit of balanced economic growth, integrated with environmental impacts and social fairness, has surpassed traditional efficiency-centric approach to become the promising trend in industrial development. This shift has led to the transition from I4.0 toward I5.0. As a critical sub-sector of logistics and supply chain management, perishable logistics, constrained by the inherent biological attributes of the products, requires more stringent technological frameworks to mitigate social and environmental influences. Currently, the discourse on sustainable perishable logistics within the I5.0 remains in the initial phase, and existing literature reviews exhibit clear limitations.

To address this gap, a systematic literature review and bibliometric analysis were conducted in this study in accordance with the PRISMA approach, based on data sourced from Scopus and Web of Science. Through a rigorous screening process, a total of 104 research articles were identified for inclusion. R programming and VOSviewer (version 1.6.19) were employed to perform a comprehensive systematic analysis, identifying influential publications, key authors, prominent journals, organizations, and contributing countries. By analyzing scient metric data through co-occurrence network analysis and thematic mapping, the key conceptual aspects of sustainability-constrained and I5.0-driven perishable product logistics were identified. Then, through a cluster analysis using co-citation and bibliographic coupling techniques, emerging research themes were identified, and corresponding future research propositions were proposed.

Summary of Research Questions

The key findings are summarized in response to each research question as follows:

With respect to RQ1, the bibliometric analysis reveals a strong geographical concentration in developing countries, particularly China (30 publications), Iran (12), and India (11). These countries exhibit an urgent necessity for industrial upgrading, which catalyzes their proactive adoption of emerging technologies. A small group of consistently active researchers dominates the field, with most contributors emerging after 2023, underscoring its novelty. The Journal of Cleaner Production and Sustainability emerge as the most influential publication venues, reflecting the interdisciplinary nature of the field across engineering, management, and environmental science.

Regarding RQ2, the co-occurrence network identifies management and optimization as the structural backbone of the field, yet their low clustering coefficients (0.157 and 0.167, respectively) reveal that both function as broad methodological umbrellas rather than cohesive research communities. Data-driven technologies such as big data and machine learning remain in the Emerging quadrant with near-zero centrality, confirming their peripheral status. Sustainability occupies a structurally balanced position with a betweenness centrality of 1843.5, reflecting its maturation as an integrative concept across research communities, while human-centricity and resilience remain absent from the keyword network.

With respect to RQ3, the cluster analysis uncovers seven emerging research themes spanning quality monitoring, sustainability management, resilient network design, I5.0-assisted monitoring, optimization frameworks, sustainable routing, and technological barriers. This thematic distribution illustrates that the discourse spans across policymaking, technical innovation, and practical implementation. Across all clusters, the sustainability pillar is relatively well-addressed, while human-centricity and resilience remain structurally underrepresented. Many studies nominally classified under I5.0 continue to operate within an I4.0 techno-centric logic. Encouragingly, there have been initial attempts in areas such as social equity, worker welfare, and data-driven technologies (IoT, digital twins, and AI), although they still do not meet the precise standards of I5.0.

Finally, regarding RQ4, sixteen future research directions are proposed across the seven clusters, with three identified as high-priority areas: “I5.0 technologies for perishable food safety management and supply chains”, “Readiness of perishable logistics and supply chain against pandemics and disruptions”, and “Resilient and flexible network design under uncertainty and potential disruptions”. Human-centricity and resilience should draw on the trajectory of discovery, definition, understanding, and maturation that sustainability has followed for decades. This may require long-term, collaborative efforts across multiple research fields (qualitative and quantitative research).

These findings are expected to serve as a solid foundation for advancing future research in perishable logistics, providing both researchers and practitioners with an overview of the I5.0 era.