Clustering Search Applied to the Periodic Vehicle Routing Problem: a Case Study in Waste Collection

The class of vehicle routing problems (VRP) involves designing the optimal set of routes for fleets of vehicles for the purpose of serving a given set of customers. A variant of the VRP is the Periodic Vehicle Routing Problem (PVRP) that consists in generating a set of routes for each day with minimal global costs and respecting all constraints of the problem. This paper presents a model for the PVRP applied to solid waste collection. To solve this problem, it is proposed an hybrid method called Clustering Search, that aims to combine metaheuristics and local search in which the search is intensified only in areas of the search space that deserve special attention. In the computational tests, a real-world case from Ponte de Lima, a Portuguese municipality is analyzed, and also some classical instances from the literature.