Studying the influence of algorithmic parameters and instance characteristics on the performance of a multiobjective algorithm using the Promethee method

Abstract

A method is proposed to analyze the effect of the algorithmic parameters and instance characteristics on the quality of a Pareto front produced by a multiobjective algorithm. This method is applied to a variable neighborhood tabu search that is used to solve a multiobjective microzone-based vehicle routing problem. Our method can accommodate many different performance indices for Pareto fronts and uses the Promethee multicriteria decision analysis method to select the best configuration based on the characteristics of the instance being solved.