Heuristic approach for solving capacitated location-routing problems

Abstract

This paper deals with the standard capacitated location-routing problem (CLRP), where two types of interdependent decisions must be made: which facilities among a number of potential ones should be operated and which vehicle routes should be built to fulfil the demand of customers from the operated facilities using a given fleet. We propose a heuristic mechanism to address the problem. The algorithm consists of four stages, namely optimally clustering customers by k-means and assessed by Dunn index, selecting the nearest depot to be opened based on Euclidean distance calculation, customers-to-depot matching and searching the optimal distribution route. This heuristic algorithm is then implemented to a dataset comprises of 20 customers and 5 potential depot facility locations.