A Comparative Study of Different Optimization Techniques for a bi-criteria Flow shop Problem

Malek Rahoual1, Mohamed-Hakim Mabed2, Clarisse Dhaenens3, El-Ghazali Talbi3
1LaMI, Université d ‘Evry Val d’Essonne, 91000 Evry – France.
2Université de technologie de Belfort Montbéliard,Laboratoire systéme et transport, Département Génie Informatique F-90010 Belfort Cedex France.
3Lift, University of Lille, Bat.M3, 59655 Villeneuve d’Ascq Cedex France.

Abstract

The resolution of workshop problems, such as the Flow Shop or the Job Shop, has great importance in industrial areas. Criteria to optimize are generally the minimization of the makespan time or the tardiness time. However, few resolution approaches take into account those different criteria simultaneously. This paper presents a comparative and progressive study of different multicriteria optimization techniques. Several strategies of selection, diversity maintaining, and hybridization will be exposed. Their performances will be compared and tested. A parallel GA model is proposed, which allows increasing the population size and the limit generations number, and leads to better results. In parallel to the work on the optimization technique, we propose here a new bi-criteria flow shop benchmark, responding to the need for common problem instances in the field of multicriteria optimization.

Keywords: Genetic Algorithm, Multicriteria optimization, Flow Shop, Sharing, Ranking, Hybridization, Local Search.