Simulation in production scheduling: scheduling flow-shops with limited buffer spaces

The work described in this paper attempts to validate the implicit assumption in traditional flow shop scheduling research that there is a buffer of infinite capacity between any two adjacent machines. The modified NEH (Nawaz, Encore and Ham) algorithm is used to generate an initial permutation schedule which is then improved by tabu search. For any given sequence, a limited equal buffer size is considered in computing job completion times. The scheduling objective is to minimize mean job flowtime. Computational results and analysis are presented. Through these simulation experiments, it was found that the improvement by tabu search can be significant and there is no need for more than 4 buffer spaces between any two adjacent machines. Future research directions are also discussed.