Abstract: To address the problem of uncertainty in facility capacity and transportation distances due to unexpected disruptions in supply networks of perishable goods, this study explores the co-optimization of low-carbon production and routing for perishables in the context of the globally open and shared physical internet (PI) environment. As an innovative logistics system characterized by connections, sharing, and adaptability, PI has great potential to enhance the resilience of perishables supply networks and reduce carbon footprints. A scenario-based two-stage stochastic programming model is developed based on the unique characteristics of perishables and the features of PI. To solve the model, a hybrid solution method integrating scenario generation and reduction with improved Benders decomposition is designed. Finally, a case study is carried out in the context of Liaoning Supply and Marketing Anbang Haider Food Supply Network to verify the feasibility of the model and the effectiveness of the algorithm. Sensitivity analysis results show that PI has significant advantages in reducing carbon footprints and enhancing supply chain resilience. Additionally, this study comprehensively analyzes the impacts of carbon emission trading mechanisms, facility capacity loss, increased transportation distances, and product lifecycle effects. This study not only enriches the theoretical foundation of low-carbon production and routing co-optimization in perishables supply networks but also provides important theoretical guidance for logistics companies in coping with uncertainty and pursuing carbon neutrality goals.

Key words: physical internet, carbon reduction, co-optimization of production and routing, uncertainty, improved Benders decomposition algorithm