Redesigning Circular Energy Supply Chains Toward Carbon Neutrality: A Multi-Objective Decision Model Integrating Mobile Manufacturing and Hybrid Intelligent Optimization

摘要

Background and Gap: Driven by global carbon neutrality targets, the photovoltaic (PV) industry has experienced explosive growth. However, the ensuing surge in end-of-life (EOL) PV modules poses severe challenges to ecological environments and resource circularity. Existing PV recycling supply chains predominantly rely on fixed processing facilities, which suffer from high transportation costs, significant carbon emissions, and delayed recycling responses in remote areas, rendering them inadequate for the efficient processing of large-scale, geographically dispersed EOL modules. Methodology: To address these issues, this study proposes a multi-objective decision-making model for redesigning circular energy supply chains by integrating mobile manufacturing facilities. The model comprehensively evaluates sustainability indicators across three dimensions: economic (minimizing total cost), environmental (minimizing carbon footprint), and social (maximizing job creation). Implementation: The research introduces mobile PV module dismantling and recycling units, enabling on-site processing of EOL modules through dynamic facility location and routing optimization. To tackle the multi-objective nature and large-scale computational demands of the model, an improved hybrid intelligent optimization algorithm combining the Non-dominated Sorting Genetic Algorithm II (NSGA-II) and Random Forest (RF) is designed. An empirical analysis is conducted based on the PV recycling network of a typical province in China. Key Findings: The results indicate that introducing mobile manufacturing facilities can significantly reduce transportation-related carbon emissions and improve local employment creation. In economic terms, although mobile facilities involve additional deployment and operation costs, they can lower total system cost under geographically dispersed recycling scenarios by reducing long-distance transportation and intermediate handling requirements. In the case study, the hybrid facility configuration shows a cost advantage over the conventional fixed-facility-only network, but such an advantage is sensitive to deployment cost, processing efficiency, and regional dispersion of end-of-life PV modules. Furthermore, the proposed hybrid intelligent algorithm improves computational efficiency by several folds compared to traditional exact algorithms when solving large-scale network optimization problems. Significance: This study not only provides an innovative theoretical framework for the design of closed-loop supply chains in the PV industry but also offers scientific decision support for governments and enterprises in formulating resource recycling policies and optimizing industrial layouts under the carbon neutrality mandate.