Abstract
Driven by global sustainable development strategies, the New Energy Vehicle (NEV) industry is expanding rapidly, and the imminent surge in End-of-Life Power Batteries (ELPB) presents significant challenges to the resilience of their reverse supply chains. Optimizing Supply Chain Resilience (SCR) requires scientifically robust evaluation methods. However, existing approaches exhibit key limitations: multi-criteria decision-making methods often lack dynamic assessment capabilities and rely heavily on subjective expert judgment, while data-driven models face high data acquisition costs and limited interpretability, especially within emerging industries. To address these gaps, this paper proposes a comprehensive SCR evaluation framework for urban ELPBs supply chains, structured around three core dimensions: robustness, flexibility, and recovery capability. The Best-Worst Method (BWM) is employed to identify the most critical indicators, which then serve as inputs to a Radial Basis Function Neural Network (RBFNN). This approach reduces the burden of data collection while enhancing model transparency. Utilizing data from 60 cities in China, a multi-modal data fusion forewarning model for SCR is developed. The model, trained on ten key indicators, including total recovery cost (C3), remanufactured battery price (C5), and number of recycling sites (S1), demonstrates high predictive accuracy, even with limited historical data. The model is further applied to forecast SCR trends over a five-year horizon in Shenzhen, Chengdu, and Shenyang, revealing how SCR dynamics align with differing urban policy orientations. Based on these insights, differentiated and targeted SCR enhancement strategies are proposed for each city, integrating controllable resilience factors with local contextual characteristics. This paper offers a scalable and interpretable tool for evaluating and strengthening SCR in the context of emerging sustainable industries.