Routing in waste collection: A simulated annealing algorithm for an Argentinean case study

Abstract

The management of the collection of Municipal Solid Waste is a complex task for local governments since it consumes a large portion of their budgets. Thus, the use of computer-aided tools to support decision-making can contribute to improve the efficiency of the system and reduce the associated costs, especially in developing countries, which usually suffer from a shortage of resources. In the present work, a simulated annealing algorithm is proposed to address the problem of designing the routes of waste collection vehicles. The proposed algorithm is compared to a commercial solver based on a mixed-integer programming formulation and two other metaheuristic algorithms, i.e., a state-of-the-art large neighborhood search and a genetic algorithm. The evaluation is carried out on both a well-known benchmark from the literature and real instances of the Argentinean city of BahȪa Blanca. The proposed algorithm was able to solve all the instances, having a performance similar to the large neighborhood procedure, while the genetic algorithm showed the worst results. The simulated annealing algorithm was also able to improve the solutions of the solver in many instances of the real dataset.

Keywords

• municipal solid waste
• waste collection
• vehicle routing problem
• simulated annealing
• large neighborhood search
• genetic algorithm
• mixed-integer programming