A closed-loop multi-agent swarm framework for synergistic location-routing optimization in municipal solid waste...

To address the challenges faced by collection and transportation systems under waste classification, such as the increasing variety of vehicle types, the growing complexity of transportation modes, and fluctuations in waste generation, this study proposes a collaborative optimization method for municipal solid waste collection and transportation that integrates multi-agent swarm modeling and bi-level programming. By constructing a bi-level location-routing model with a closed loop feedback mechanism, and designing a hybrid algorithm combining genetic algorithm and large neighborhood search to efficiently solve the capacitated vehicle routing problem with time windows, the method achieves synergistic optimization. Results show that the proposed retrofit plan for transfer stations reduces total costs by 10% compared to the original system, while the optimal integrated cost among new construction plans is achieved when four transfer stations are built. Furthermore, scenario simulations analyze the economic and environmental impacts of different operational factors. This study not only provides methodological tools for dynamic multi-vehicle collection and transportation systems but also establishes a multi-agent simulation framework that can serve as the core of a smart waste management platform, offering practical significance for enhancing the efficiency and sustainability of urban solid waste systems.