Optimization by simulation in manufacturing flow control

A class of piecewise-deterministic control problems typically associated with manufacturing flow-control models is considered. The authors consider different approaches for the definition of a sub-optimal control policy. In particular, the method consisting of optimizing a policy in a class indexed over a given parameter set is addressed. This approach transforms a stochastic control problem into a parameter design problem. It is shown how these models can be used to represent preventive maintenance and/or setup costs in manufacturing flow-control models. The possibility of using an optimization-by-simulation approach to solve the policy-design problem is discussed