Sampling-Based Genetic Algorithms for the Bi-Objective Stochastic Covering Tour Problem

Abstract

The paper investigates a sampling-based extension of the NSGA-II algorithm, applied to the solution of a bi-objective stochastic covering tour problem. The proposed extension uses variable samples for gradually improving approximations to the Pareto front. The approach is evaluated on a test benchmark for a humanitarian logistics application with data from Senegal. Comparisons to alternative solution techniques, in particular also to the exact solution of the deterministic counterpart problem based on a fixed sample, show the superiority of our approach.