Randomised Variable Neighbourhood Search for Multi Objective Optimisation

TL;DR

This paper investigates the effectiveness of various neighborhood search operators in multi-objective flow shop scheduling, demonstrating that hybridising these operators with a randomized variable neighborhood search improves the identification of Pareto optimal solutions.

Contribution

It introduces a hybridized randomized variable neighborhood search method tailored for multi-objective scheduling, highlighting the impact of operator interdependencies on solution quality.

Findings

No single neighborhood operator finds all Pareto optimal solutions.

Hybridized approach significantly improves solution quality.

Statistical tests confirm the effectiveness of the hybrid method.

Abstract

Various local search approaches have recently been applied to machine scheduling problems under multiple objectives. Their foremost consideration is the identification of the set of Pareto optimal alternatives. An important aspect of successfully solving these problems lies in the definition of an appropriate neighbourhood structure. Unclear in this context remains, how interdependencies within the fitness landscape affect the resolution of the problem. The paper presents a study of neighbourhood search operators for multiple objective flow shop scheduling. Experiments have been carried out with twelve different combinations of criteria. To derive exact conclusions, small problem instances, for which the optimal solutions are known, have been chosen. Statistical tests show that no single neighbourhood operator is able to equally identify all Pareto optimal alternatives. Significant improvements however have been obtained by hybridising the solution algorithm using a randomised variable neighbourhood search technique.