A bi-objective metaheuristic for the problem of sequencing repetitive jobs in a flowshop heijunka production setting

Abstract

In this paper, a sequencing problem in a flowshop environment was studied, taking into account non-unitary demands for jobs, seeking to minimize the total completion time of all jobs ( ) while maintaining an adequate production mix throughout the sequence. A procedure to solve the problem was proposed: a bi-objective metaheuristic based on simulated annealing and tabu search. The proposed bi-objective algorithm generated a set of solutions that represents an approximation to the Pareto optimal frontier. An experimental work was developed to evaluate the performance of the technique used, resorting to the set of instances corresponding to a problem of industrial scale, obtaining a satisfactory preliminary performance both in terms of the quality of the sets of solutions obtained, as well as the computational effort required. Through the adopted solution approach, applied to the analyzed case of study, sequences that represent an adequate compromise between production lead times and the establishment of a uniform work flow in the facility can be selected.