Grouping genetic algorithms for a bi-objective inventory routing problem

Abstract

Inventory routing problems (IRPs) is one of the versions of vehicle routing problems (VRPs), which retains the attention of the researchers. The main idea is to coordinate the distribution plan with the inventory management in a same model. The problem studied is to determine the multi-tours of a homogeneous fleet of vehicles covering a set of sales-points and minimising the distribution and inventory cost per hour. Of course no stock-outs are acceptable at the sales-points. In this paper, we analyse this problem like a bi-objective inventory routing problem in which the transportation cost and the delivery cost are considered separately. Two approaches are proposed to approximate the Pareto front of this bi-objective problem. Both methods are an adaptation of the hybrid grouping genetic algorithm (HGGA) that we proposed for the single objective problem in which a grouping genetic algorithm is combined with a local search. Computational experiments are reported using eight instances for four groups: (25, 50,100 and 200 sales points).