Improving the design of stochastic production lines: An approach using perturbation analysis

With the recent increase in expense and specialization of equipment, capacity decisions have taken on greater significance. As a result, companies are in need of a better understanding of the investment tradeoffs. In this paper, we examine the problem of determining the most profitable capacity configuration for a production line modelled as a series of single-server stations. In the context of a constant work-in-process (CONWIP) control system, an algorithm is developed for solving the general problem using a single-run simulation procedure. Various market structures are examined and sensitivity analysis is performed on the cost of capacity, quality and the amount of work-in-process allowed in the system.