Anais do Simpósio Brasileiro de Pesquisa Operacional
Anais do Simpósio Brasileiro de Pesquisa Operacional

This paper addresses the backlog sequencing problem in a flow shop controlled by a ConWiP production control model when processing times are random, and the objective is to minimize the expected makespan. For this problem, we evaluated the performance of two constructive heuristics: a deterministic heuristic applied to mean processing times and a heuristic designed specifically for the stochastic problem. We use simulation to estimate the expected makespan. In general, stochastic heuristic perform better than deterministic heuristic applied to mean processing times. However, this difference does not justify the computational cost of stochastic heuristic, at least within the limits of our experiments. Therefore, a more practical way to deal the stochastic problem is to resolve the deterministic counterpart, since it provides good sequences in reasonable computational time.