Abstract: To address the issue of local failures in the fresh agricultural product supply chain during emergencies, complex network robustness theory was introduced to optimize the emergency supply logistics network for fresh agricultural products. A robustness evaluation index system for node physical damage resistance, topological connectivity and timeliness robustness, supply-demand coordinated response robustness, and macro mechanism adaptability robustness was constructed. A game composite weighting method based on inner product space was used to coordinate subjective and objective weight differences, and a fuzzy comprehensive evaluation model was used to quantitatively evaluate the initial state of the emergency supply logistics network in the megacity W. On this basis, a mixed-integer programming model was constructed to minimize operating costs and maximize system robustness. Results showed that the initial comprehensive robustness score for the emergency supply logistics network in W was 2.93, with shortcomings in topological connectivity, node redundancy, and path reconstruction; after network topology reconstruction, the comprehensive robustness score increased to 4.18, an increase of 42.7% compared with the initial state, and structure redundancy of the network and risk response capacity were improved. The research results could provide quantitative references for optimizing and enhancing the robustness of urban fresh agricultural product emergency supply logistics networks under emergency scenarios.

Key words: emergencies, fresh food supply, robustness measurement, game combination weighting, dual-objective programming