Production control in a failure-prone manufacturing network using discrete event simulation and automated response surface methodology

In this paper, a system consisting of a network of machines with random breakdown and repair times is considered. The machines in this system can be in one of four states: operational, in repair, starved, and blocked. Failure and repair times of the machines are exponentially distributed. Previous research on multi-machine failure-prone manufacturing systems (FPMS) has focused on systems consisting of machines in series or in parallel. This paper considers a network of machines with relationship constraints. Additionally, the system under study models work in process for multiple products, intermediate and final buffers and one type of final product. The demand rate for the final commodity is constant and unmet demand is either backlogged or lost. The objective of this control problem is to find the production rates and policies of the different machines so as to minimize the long run average inventory and backlog cost. The applied control policy is the hedging point policy that is determined by factors representing the level of buffer inventory for each machine. Obtaining analytical solutions is generally impossible for such complex systems. To simultaneously control the production rates of the machines we have therefore developed a method based on a combination of stochastic optimal control theory, discrete event simulation, experimental design and automated response surface methodology (RSM). The application of an automated RSM for Network FPMS is another contribution of this paper. The model can be extended easily to systems with age-dependent failure rates, a preventive repair maintenance policy and non-exponentially distributed up and down times.