An improved adaptive large neighborhood search algorithm to solve a bi-level medical waste location-routing problem with infection control

Abstract

The rapid urbanization and population growth in major cities worldwide have led to a significant increase in medical waste generation, often containing infectious materials that require stringent handling protocols. To address the complexity of vehicle allocation and routing in this context, efficient planning methods are essential. This study introduces a comprehensive approach to the medical waste location-routing problem, incorporating multiple practical constraints such as vehicle capacity, hospital classification, infection risks, and time-window restrictions. Our novel solution integrates an exact algorithm for optimizing transfer center locations and collection routes at the upper level, combined with an improved adaptive large neighborhood search (IALNS) for routing optimization at the lower level. The IALNS leverages enhanced neighborhood exploration techniques and Pareto ranking with reward adjustment method to balance total cost and infection risk. Simulations based on real-world data from Chengdu, China, validate the effectiveness of the proposed method. Additionally, comparisons with Gurobi and other representative metaheuristic algorithms on randomly generated instances and benchmark datasets further demonstrate the superior efficiency and solution quality of the IALNS algorithm. This research provides government authorities with a practical and robust strategy for transporting infectious medical waste, enhancing both operational efficiency and public health safety.