Hybrid Iterative Local Search Heuristic with a Multiple Criteria Approach for the Vehicle Routing Problem

Abstract

This paper presents a Hybrid Iterative Local Search Heuristic and its framework, whose aim lies on helping to escape from local optima when a construction heuristic, for the VRP, has been trapped. The approach was mainly inspired by basic and modified versions of related successfully applied heuristics such as Variable Neighborhood Search (VNS) and Granular Local Search (GLS). Differently to a great deal of local search heuristics revised, which mainly consider a single decision criterion, multiple optimization criteria are considered all along the local search and a Multiple Criteria Threshold has been proposed allowing defining which arcs must be included in a candidate list to explore, this considerably reduces the search area and has a major incidence in the satisfaction of the clients. The proposal is enriched with a good literature review, taking into account some of the gaps and achievements of the states of the art and practice.