Comparing Single and Multiobjective Evolutionary Approaches to the Inventory and Transportation Problem

TL;DR

This paper compares single and multiobjective evolutionary algorithms for the Inventory and Transportation Problem in retail, concluding that single objective approaches are generally more effective for the studied case.

Contribution

It evaluates the effectiveness of multiobjective versus single objective evolutionary approaches for ITP, providing insights into their relative performance.

Findings

Single objective approach is generally preferable for ITP in the studied case.

Using Clarke & Wright's algorithm as a seed improves VRP solutions.

Multiobjective approach does not outperform single objective in this context.

Abstract

EVITA, standing for Evolutionary Inventory and Transportation Algorithm, is a two-level methodology designed to address the Inventory and Transportation Problem (ITP) in retail chains. The top level uses an evolutionary algorithm to obtain delivery patterns for each shop on a weekly basis so as to minimise the inventory costs, while the bottom level solves the Vehicle Routing Problem (VRP) for every day in order to obtain the minimum transport costs associated to a particular set of patterns. The aim of this paper is to investigate whether a multiobjective approach to this problem can yield any advantage over the previously used single objective approach. The analysis performed allows us to conclude that this is not the case and that the single objective approach is in gene- ral preferable for the ITP in the case studied. A further conclusion is that it is useful to employ a classical algorithm such as Clarke & Wright's as the seed for other metaheuristics like local search or tabu search in order to provide good results for the Vehicle Routing Problem.