On dominance-based multiobjective local search: design, implementation and experimental analysis on scheduling and traveling salesman problems |
| |
Authors: | Arnaud Liefooghe Jérémie Humeau Salma Mesmoudi Laetitia Jourdan El-Ghazali Talbi |
| |
Institution: | 1. Laboratoire d??Informatique Fondamentale de Lille (LIFL), UMR CNRS 8022, Universit?? Lille 1 INRIA Lille-Nord Europe, Cit?? scientifique, Bat. M3, 59655, Villeneuve d??Ascq cedex, France 2. D??partement IA, ??cole des Mines de Douai, 941 rue Charles Bourseul, BP 10838, 59508, Douai, France 3. Laboratoire d??Imagerie Fonctionnelle (LIF), UMR_S 678 Inserm?CUPMC, CHU Piti??-Salp??tri??re, 91 boulevard de l??H?pital, 75634, Paris cedex 13, France 4. King Saud University, Riyadh, Saudi Arabia
|
| |
Abstract: | This paper discusses simple local search approaches for approximating the efficient set of multiobjective combinatorial optimization
problems. We focus on algorithms defined by a neighborhood structure and a dominance relation that iteratively improve an
archive of nondominated solutions. Such methods are referred to as dominance-based multiobjective local search. We first provide
a concise overview of existing algorithms, and we propose a model trying to unify them through a fine-grained decomposition.
The main problem-independent search components of dominance relation, solution selection, neighborhood exploration and archiving
are largely discussed. Then, a number of state-of-the-art and original strategies are experimented on solving a permutation
flowshop scheduling problem and a traveling salesman problem, both on a two- and a three-objective formulation. Experimental
results and a statistical comparison are reported in the paper, and some directions for future research are highlighted. |
| |
Keywords: | |
本文献已被 SpringerLink 等数据库收录! |
|