Resilience assessment and prediction of fresh agricultural product supply chains using variable fuzzy recognition and machine learning

Fresh agricultural products are highly perishable and sensitive to transportation, storage, and handling conditions, which makes their supply chains particularly vulnerable to disruptions. Such disruptions can interrupt product flows and reduce both quality and availability. Traditional risk analysis methods mainly focus on uncertainty and the probability of adverse events, but they do not adequately capture the ability of supply chains to recover and adapt following disruptions. To address this limitation, this study develops a resilience evaluation and prediction framework for fresh agricultural product supply chains (FAPSCs). The framework is structured around three dimensions: production, transportation, and consumption. A hybrid modeling approach is proposed by integrating variable fuzzy set theory with machine learning. The variable fuzzy set model is used for dynamic resilience assessment, while a convolutional neural network (CNN) is employed to capture temporal dynamics and perform prediction. Indicator weights are determined by combining the Analytic Hierarchy Process (AHP) and the CRITIC method to improve robustness and interpretability. Using panel data from 31 provinces in China from 2009 to 2023, the results show that the proposed framework effectively distinguishes different resilience levels across regions. The CNN model captures temporal variations, including peak and off-peak fluctuations, and outperforms baseline models such as Decision Tree, Random Forest, and Linear Regression, particularly in modeling nonlinear patterns and high-resilience cases. Overall, the proposed approach demonstrates strong robustness under limited or noisy data conditions and provides practical support for early warning, risk mitigation, and decision making in FAPSC management.