Local search heuristics for sectoring routing in a household waste collection context

Abstract

This paper addresses the problem of residential waste collection, as a real life application of a sectoring-arc routing problem (SARP). Tactical decisions comprise the partition of the service territory into a number of sectors so that each sector can be covered by a set of vehicle trips. In addition, operational decisions involving the design of the vehicle trips that minimize total routing time are to be made. Apart from supporting good vehicle trips, sectors should also be planned such that the workload time imbalance as well as the number of connected components are minimized. These later try to promote service areas (sectors) geographically concentrated and grouped into delimited regions. We propose two local search methods: a hill climbing and a tabu search based heuristic. A constructive heuristic for obtaining an initial solution is also suggested. By means of a normalized weighted sum of criteria for evaluating solutions, the local search heuristics were tailored to improve the features of the initial solution. The algorithms are tested on random instances and also on real life based instances. The results show that the proposed local search methods are an efficient way of obtaining good quality solutions to implement in practice. Results also highlight that the proposed function for evaluating solutions during the search phase plays an essential role.