Optimal input control for job-shop type production systems using inverse queueing network analysis

This paper describes an input controlling problem of the job-shop production type system where multiple model of products are produced repeatedly. The authors present an efficient method to obtain the desirable level of work-in-process which achieves the required throughputs and approximately minimizes the residence time of each product. The method uses a closed queueing network model and a simulated annealing algorithm alternately to achieve the objectives of the problem. The computational results of a sample problem on the hypothetical semiconductor manufacturing line are shown and the significance of the method presented is discussed.