Conclusions

Abstract

This thesis has addressed a vital component in the domain of Logistics and Distribution Management, viz., the Travelling Salesman Problem (TSP) and the Vehicle Routing Problem (VRP). As is abundantly clear by now, the TSP and VRP and its variants in real-life are not “solution-friendly”. Belonging as they do, to the class of NP-hard problems, they are not amenable to the determination of optimal solutions. While the choice of methodology inevitably falls on meta-heuristics, the inherent characteristics of the problems are analysed and polynomial algorithms are devised, at least, to find good feeder solutions to meta-heuristics. The choice of construction heuristics to find initial solutions has contributed significantly in the reduction of computational effort and for better deployment of the meta-heuristics. These are born out of the results obtained over several datasets for all the variants of TSP and VRP considered in this study. One of the key contributions of this study is the unified heuristics that are developed for selected complex variants of TSP and VRP and the enhancement of their capability to solve the simpler versions of the problem as well.