Authors: S. Hamed Hendizadeh, Tarek Y. ElMekkawy, G. Gary Wang
Addresses: Department of Mechanical and Manufacturing Engineering, University of Manitoba, Winnipeg, MB R3T 5V6, Canada. ' Department of Mechanical and Manufacturing Engineering, University of Manitoba, Winnipeg, MB R3T 5V6, Canada. ' Department of Mechanical and Manufacturing Engineering, University of Manitoba, Winnipeg, MB R3T 5V6, Canada
Abstract: The paper considers a flowshop scheduling problem of a manufacturing cell that contains families of jobs whose setup times are dependent on the manufacturing sequence of the families. Two objectives, namely the makespan and total flow time, have been considered simultaneously in this work. Since minimisation of each of these two objectives is an Np-Hard problem, a Multi-Objective Genetic Algorithm (MOGA) has been proposed to deal with the bi-criteria optimisation problem. The performance of the proposed MOGA is compared with the makespan and total flow time lower bounds. The proposed MOGA obtained solutions that only deviate by an average of 1% from the lower bounds. Future research will develop more efficient lower bounds for the total flow time and also compare the proposed method with other multiobjective meta-heuristics. [Received on 6 February 2007; Revised 6 June 2007; Accepted 16 June 2007]
Keywords: bi-criteria scheduling; cellular manufacturing; manufacturing cells; sequence-dependent setups; setup times; flowshop scheduling; pareto-optimal frontier; multi-objective genetic algorithms; MOGA; makespan; total flow time; lower bound; branch and bound; B&B; metaheuristics.
European Journal of Industrial Engineering, 2007 Vol.1 No.4, pp.391 - 413
Published online: 14 Oct 2007 *Full-text access for editors Access for subscribers Purchase this article Comment on this article