An interactive optimization approach to a realwor

Authors: Izaskun Ibarbia Alexander Mendiburu Maria Santos Jose A Lozano

Publish Date: 2011/04/26

Volume: 36, Issue: 3, Pages: 721-734

PDF Link

Abstract

In this paper we deal with a realworld routing problem that consists of finding the best route for an oceanographic campaign The planning task involves setting up the initial route and managing the route variations along the way Bearing in mind the idea of building an experimental tool to support the decision making process of the expert when planning the route two different approaches to solve the problem were developed First of all we assumed that the problem was well characterized by the initial description of the expert using as fitness function the time employed in the route and applying classical local search optimization heuristics The difficulty of a precise constraint definition and the infeasibility of mathematically describing all the evaluation criteria of the expert led us towards interactive optimization as a way to introduce the knowledge of the expert According to the results this last approach was able to provide satisfactory solutions

Keywords: