The agricultural product supply chain is a network of interconnected individuals and organizations involved in the production, transportation, distribution, and consumption of agricultural products (Zhao et al., 2023). The goal of agricultural product supply chain management is to ensure the efficient, safe, and sustainable flow of agricultural products from production to consumption while meeting consumers’ needs and expectations (Fang and Ge, 2023). Early research in the field of agricultural product supply chains focused on supply chain organizational models suitable for the market and the characteristics of agricultural products (Routroy and Behera, 2017; Syahruddin and Kalchschmidt, 2012), optimal resource allocation from a supply chain integration perspective (Bian and Tian, 2022; Cui, 2021), supply chain risk management (Behzadi et al., 2018), and supply chain performance evaluation (Kamble et al., 2020). Since 2020, the impact of the global COVID-19 pandemic has led to frequent production disruptions and transportation obstacles (Raihan and Himu, 2023), increasing supply chain interruptions and market uncertainty, and consequently, the incidence of supply chain breakages has risen. Against this backdrop, research on supply chain resilience has become a current hot topic (Roberta Pereira et al., 2014; Scholten and Schilder, 2015; Tukamuhabwa et al., 2015). Its practical value is guaranteed by considering the perishability of agricultural products, circulation losses, and the characteristics of agricultural product supply chain forms under different operational models. Some scholars have also focused on the interaction between the agricultural product supply chain and the environment and society, such as the environmental impact of production processes, carbon emissions during circulation, and the impact of cross-sectoral technological integration on the agricultural product supply chain. This has led to the evolution and improvement of the theoretical system of sustainable development for agricultural product supply chains. Compared to crop-based agricultural products, the management of PSC exhibits certain peculiarities. The upstream elements such as feed, vaccines, veterinary drugs, and piglets; the midstream elements like pig farming and cultivation; and the downstream elements including slaughtering, processing, and sales channels differ significantly from efficiency-oriented agricultural products (such as grains, fruits, and vegetables). However, current research often targets broad categories of agricultural products (Guohua, 2013; Yan et al., 2020), with specialized studies on PSC being relatively rare. This has resulted in overly macro-level research outcomes that lack guidance on the optimization of PSC.

By reviewing the research outcomes of agricultural and pig supply chains, it is evident that in recent years, resilience and sustainability have become the main directions for supply chain optimization. Some scholars have recognized the important relationship between supply chain resilience and sustainability, but there is still a lack of comprehensive research on their synergy. Moreover, research on the synergy of resilience and sustainability within PSC remains largely unexplored.

As supply chain management research evolves toward networked governance, collaboration theory serves as a fundamental perspective for tackling systemic synergy challenges (Touboulic and Walker, 2015b). Supply chain collaboration is the process of forming a strategic partnership among supply chain members to facilitate deep information exchange, resource integration, goal alignment, and joint decision-making to achieve a shared value-added process (Cao and Zhang, 2011). The fundamental mechanisms of collaborative innovation and benefit sharing create the theoretical foundation for achieving RSS in the PSC. Collaborative innovation necessitates that members jointly invest complementary resources and develop systemic solutions through their collective actions (Soosay et al., 2008). At the resilience level, supply chain collaboration improves visibility, predictive abilities, and risk response speed through information sharing and joint planning (Scholten et al., 2019). At the sustainability level, it encourages the collaborative use of eco-friendly technologies to share risks (Beske and Seuring, 2014). Benefit sharing maintains collaborative viability, fair profit-distribution mechanisms encourage vulnerable entities to engage long-term in resilience-building (Hingley, 2005), while scientifically crafted sharing models convert sustainability initiatives into economic incentives, mutually reinforcing RSS (Touboulic and Walker, 2015a).

In summary, supply chain collaboration theory offers a strong theoretical foundation for understanding and achieving the RSS. However, when examining China’s complex and nationally crucial PSC, which is vital to both the national economy and people’s livelihoods, current research has not yet developed a systematic, clear, and locally contextualized framework for identifying the key factors that drive or constrain resilience and sustainability within this system.

The resilience of agricultural product supply chains is defined as the ability of agricultural food supply chain stakeholders to ensure an acceptable, adequate, and stable food supply at the required time and place by accurately predicting disruptions and using strategies that delay impacts, aid recovery, and allow for cumulative learning after disruptions (Montanyà and Amat, 2023). Scholars have studied the key factors affecting the resilience of agricultural product supply chains under different conditions and dimensions. From a macro-management perspective, Coopmans et al. (2021) suggest that the resilience of agricultural product supply chains is influenced by diversity, flexibility, openness, and self-organization (Coopmans et al., 2021; Ivanov and Das, 2020) identify production inventory dynamics, customer performance, financial performance, and delivery time performance as determinants of supply chain elasticity (Ivanov and Das, 2020). From the disruption recovery perspective, Dixit et al. quantify supply chain recovery capability through unmet demand capacity and total transportation cost control post-disaster (Dixit et al., 2016). Mishra et al. (2022) indicate that training company personnel in disruption awareness and management is a critical factor (Mishra et al., 2022). Establishing rapid response mechanisms, robust supply chain information systems, highly coordinated partnerships, and pre-planning for post-disaster reconstruction are also key elements for enhancing recovery capabilities. For PSC, Pfeifer et al. (2022) investigate the resilience strategies of European organic pig producers in response to economic, legislative, labor, and climate-related shocks. Analyzing narratives from 18 producers, it identifies three strategies: efficiency-based, nutrient substitution, and farm diversification. Ntakiyisumba et al. confirm the widespread prevalence of Salmonella in the South Korean PSC and identify critical factors that contribute towards the Salmonella contamination of pork carcasses at slaughter (Ntakiyisumba et al., 2023).

Research on the resilience of agricultural product supply chains mainly focuses on operational robustness, flexibility, and post-disruption recovery capabilities. However, existing studies have not comprehensively considered characteristic risks in PSC, such as the hog cycle, environmental regulations, and health safety incidents. There is still no complete assessment system for the resilience of PSC.

The sustainability of agricultural product supply chains influences food safety, environmental protection, and social welfare. Its key influencing factors have been increasingly discussed by scholars. Researchers have explored the impact of technology diffusion on agricultural product supply chains. Nandi et al. believe that redesigning supply chains through innovative technologies such as blockchain can better manage food safety (Nandi et al., 2021). Fesharaki and Safarzadeh (2022) found that the integration of refrigeration systems and the adoption of IoT (Internet of Things) ecosystems are crucial for overcoming perishability limitations and ensuring equitable distribution of crops. Su et al. (2023) also suggest that utilizing high-tech technologies and using recyclable materials play a vital role in improving environmental sustainability. Social factors have also been widely discussed. Chakrabarty and Nandi (2022) argue that risks related to consumer preferences and market demand are significant constraints on sustainability. Environmental factors are also considered, including the level of sustainable environmental policy support, the state of the natural environment, the ability to reuse organic waste, and ecological compensation.

The sustainability of PSC currently focuses on two main directions. First is the continuous improvement of pig product quality. Scholars have conducted in-depth research on vaccines, breeding, and feed, with current attention points on the traceability of pig products. Second, the coordinated development of PSC and the environment. However, research outcomes in this area are relatively scarce. There is a lack of necessary theoretical support for issues such as the environmental impact of pig farming, the treatment of organic pig waste, and clean emissions in the supply chain.

Through a literature review, the factors that influence the synergy between resilience and sustainability in PSC are summarized. Different sub-factors are organized under broader categories based on their meanings, and sub-factors with the same meaning are consolidated, eliminating any duplicates. Based on these steps, this paper proposes a system for the influencing factors of the synergy between resilience and sustainability in PSC, which includes six main categories: (i) Response Capacity, (ii) Adaptive Capacity, (iii) Recovery Capacity, (iv) Environmental Impact, (v) Economic Impact, and (vi) Social Impact. Detailed descriptions of the sub-factors under each main category are provided in Table 1.

The Delphi method involves sending questionnaires to relevant experts to gather their opinions after the indicator system’s initial construction. Based on the feedback received, the indicator system is refined. After the formation of the first assessment indicator system, the decision on the retention of each indicator entry is based on the results of expert scoring. To ensure the study’s scientific rigor, objectivity, and data quality, this paper employs a generalized Likert scale, assigning values ranging from very important (5 points) to very unimportant (1 point). Experts are asked to rate the importance of each indicator.

For indicator screening, an indicator is considered for exclusion if it meets any of the following criteria: mean score below 3.5, coefficient of variation greater than 0.25, or full score frequency lower than 0.3. To ensure critical indicators are retained, if an indicator fails to meet only one criterion, its retention or exclusion will be determined through a panel discussion. The first round questionnaire is distributed for statistical analysis. Based on the results, a revised questionnaire is developed for the second round. This iterative process continues until expert opinions converge, resulting in the final evaluation index system.

This paper combines Fuzzy Theory with the Decision-Making Trial and Evaluation Laboratory (DEMATEL) method to construct a clear map of interactive relationships among influencing factors. This map illustrates the influence and dependence of each factor within the network. Subsequently, by computing a comprehensive influence matrix, it simplifies the weight assignment problem effectively, reducing the complexity of pairwise comparison processes in the Analytic Network Process (ANP) method. Considering the complex and dynamic nature of the PSC environment, the diversity of influencing factors, and the incompleteness of data, the Fuzzy Decision-Making ANP (Fuzzy DANP) method can provide more precise, comprehensive, and adaptive decision support. The key implementation steps are as follows:

After ANP obtains the indicators’ weights, this paper uses the fuzzy comprehensive evaluation method to evaluate the resilience and sustainability synergy of the pig supply chain in different regions of China.

After creating the initial influencing indicator system for RSS in PSC (as outlined in Table 1), the Delphi method was employed to finalize the RSS impact indicator system. To achieve this, five experts were selected to form a panel, as detailed in Table 2.

Data were collected through a questionnaire, and two rounds of expert consultations were conducted to achieve consensus among the participants.

The results of the first round of expert consultation, as shown in Table 3, reveal that the average rating for each first-level indicator is above 3.5 points. Additionally, the coefficient of variation (CV) is less than 0.25, and the frequency of Full Scores exceeds 0.3. These findings indicate a strong agreement among experts regarding the importance of the indicators. Since the experts did not suggest any changes to the first-level indicators, no modifications were made during this round.

In the results of the consultation regarding secondary indicators, the following indicators did not meet the retention criteria and have been removed: Managerial Qualities, Sales Adaptability, Cost Fluctuation Control Capability, and Collaboration Efficiency Among Entities. The compiled indicators were redistributed to the experts for a second round of consultation, and the results were calculated to meet the retention criteria, as shown in Table 4.

In summary, after two rounds of the Delphi method, this paper establishes an indicator system for RSS influencing factors in PSC, which includes 6 primary indicators and 21 secondary indicators, as shown in Table 5.

In Step 1, based on the final Influencing indicator system, the influence of one factor on another is divided into five scales: no influence (0), very low influence (1), low influence (2), high influence (3), and very high influence (4). The fuzzy linguistic scale is shown in Table 6.

In Step 2, five experts rated the relationships between factors based on their experience. According to Steps 3–5, the crisp total-relation matrix (defuzzified total influence matrix) is obtained, and the influence degree and influenced degree of each indicator are calculated using the equations in Step 6. This helps determine their center degree and cause degree. Based on the positive or negative values of D-R, the influencing indicators are categorized into two main types: cause indicators and effect indicators, as shown in Table 7.

Ranking the center degree of cause indicators, it can be seen that enterprise organizational capability (F2), logistics network adaptability (F7), and Trust relationships among entities (F19) are the top three in importance. When ranking cause indicators by their cause degree, enterprise organizational capability (F2), Supply chain digitalization level (F4), and Learning and continuous improvement mechanisms (F12) are identified as the most direct influencing indicators. Effect indicators are more affected by other indicators, indicating they are susceptible to influence within the system.

Based on the data in Table 7, a causal quadrant diagram of the key influencing indicators for RSS in PSC is plotted, with center degree on the horizontal axis and cause degree on the vertical axis, as shown in Figure 2.

It can be observed that:

In Step 7, the unweighted supermatrix of the Analytic Network Process (ANP) method is created from the defuzzified total influence matrix obtained through the Fuzzy-Dematel approach. In Step 8, this unweighted supermatrix is standardized to produce the weighted supermatrix. Step 9 involves applying the power iteration method to the weighted supermatrix until it converges to a stable limit supermatrix. From this stable matrix, the weights of each influencing factor are derived. Finally, by combining the centrality and ANP rankings, the comprehensive ranking results are calculated, as shown in Table 8.

Based on the comprehensive ranking, the top 7 influencing factors identified as critical for the synergistic development of resilience and sustainability in the PSC are: Enterprise organizational capability (F2), Supply chain environmental interaction across the lifecycle (F13), Trust relationships among entities (F19), Logistics network adaptability (F7), Learning and continuous improvement mechanisms (F12), Supply chain digitalization level (F4), and Technological innovation capability (F8).

Based on the level of economic development, pig industry layout and supply chain structure differences in eastern, central and western China, this paper selects Zhejiang Province (Eastern China), Henan Province (Central China) and Sichuan Province (Western China) as the research objects to analyze the RSS of PSC.

Zhejiang Province (Eastern China) represents a typical exogenous consumption-driven PSC model. The province is limited by land resources, and more than 60% of pigs need to be transferred out. However, the downstream of its PSC has significant advantages, relying on the consumption power of urban agglomerations and digital infrastructure to form downstream clusters centered on deep processing and cold chain logistics. The core challenge lies in the lack of resilience in the upstream of the supply chain.

Henan Province (Central China) has built a leading endogenous production-led PSC model. Relying on the resource endowment of the “grain silo of the Central Plains,” its supply chain is characterized by “large-scale production and full-chain integration,” with a self-sufficiency rate of pork reaching 120%, and the annual net outward transfer volume ranking among the top three in the country. Among them, Muyuan Group has an annual production capacity of more than 30 million heads, driving the intensive production of nearly 70% of farms in the province; Shuanghui Group has an annual slaughtering and processing capacity of 15 million heads, and through vertical integration, it has reduced the loss rate of the cold chain to less than 3%. Core challenges center on capacity concentration risks and environmental pressures.

Sichuan Province (Western China) has developed a unique ecologically sensitive composite PSC model. The province integrates ecological free-range farming in the highlands with intensive production in the Chengdu Plain, with a breeding cycle system covering 82 counties and the highest ecological carrying capacity in the West. The current development urgently needs to seek a balance between ecological sensitivity and industrial sustainability, both in terms of resolving restrictions on the expansion of the no-farming zone and the significant risk of logistical disruption due to the geographic conditions of the seismic zone.

This paper invited ten experts to assess the evaluation levels of 21 factors across three provinces, ultimately compiling the provincial RSS evaluation results in Tables 9–11.

The results show that Zhejiang Province (Eastern China) scored 4.15 points for the RSS in PSC, with a rating of “high (lower)”; Henan Province (Central China) scored 3.67 points, with a rating of “relatively high (upper)”; Sichuan Province (Western China) scored 3.17, with a rating of “relatively high (lower).” This result marks a significant departure from the traditional view that “capacity size determines stability,” where “Central China” is ranked higher than “Eastern China,” which is in turn ranked higher than “Western China.” The primary reason for this change is that digitalization on the consumer side enhances resilience beyond the conventional benefits associated with the capacity size. Specifically, the KIFs related to the RSS in Zhejiang, including F4, F5, F7, F8, F9, F12, and F21, all received scores exceeding 4.5 points in the fuzzy composite evaluation, indicating an advanced level of performance and forming a notable competitive advantage.

Moreover, the analysis finds that resilience and sustainable development face a profound conflict. Downstream digitization capacity (F4 of 4.7) in Zhejiang Province improves cold chain responsiveness. At the same time, land resource constraints weaken upstream manure resource utilization capacity (F4 of 3.8), creating a dumbbell imbalance of “high digital resilience-low environmental sustainability.” In Henan Province, the elasticity of large-scale production (F6, 4.1) guarantees a stable supply. Still, high farming density leads to environmental overload (F13, 2.4), reflecting the structural contradiction of “strong production elasticity-weak ecological sustainability.” The highland eco-agriculture model (F15, 4.1) maintains environmental sustainability in Sichuan Province. Still, the logistical vulnerability of the seismic zone (F7, 2.3) threatens the continuity of the supply chain, highlighting the dilemma between ecological protection and operational resilience.

While the differences in the three provinces’ performance on the KIFs reflect differences in regional development priorities, their core set of factors remains consistent: F2 ranks in the top three in all three provinces, confirming its robustness in the first place in the ANP weights; F13 has a significant weight in the ecologically sensitive Sichuan province; and F19 scores the highest in Henan, echoing the strong drive shown in the causal quadrant diagram showing strong drivers. Thus, regional differences did not change the relative importance of the KIFs and verified the stability.