A Performance Study for the Multi-objective Ant Colony Optimization Algorithms on the Job Shop Scheduling Problem

Abstract

Attached

M ost o f the research on jo b shop scheduling problem are concerned w ith m inim ization o f a single objective. How ever, the real w orld applications o f jo b shop scheduling problem s are involved in optim izing m ultiple objectives. T herefore, in recent years ant colony optim ization algorithm s have been proposed to solve jo b shop scheduling problem s w ith m ultiple objectives. In this paper, som e recent m ulti-objective ant colony optim ization algorithm s are review ed and are applied to the jo b shop scheduling problem by considering tw o, three and four objectives. A lso in this study, four criteria: m akespan, m ean flow tim e, m ean tardiness and m ean m achine idle tim e are considered fo r sim ultaneous optim ization. Two types o f m odels are used by changing the num ber o f ants in a colony and each m ulti-objective ant colony optim ization algorithm is applied to sixteen benchm ark problem instances o f up to 2 0 jo b s x 5 m achines, for evaluating the perform ances o f these algorithm s. A detailed analysis is perform ed using the perform ance indicators, and the experim ental results have show n that the perform ance o f som e m ulti-objective ant colony optim ization algorithm s depend on the num ber o f objectives and the num ber o f ants.