Costing a finite minimal repair replacement policy

In this paper an integral equation technique is used to evaluate the expected cost for the period (0,t] of a policy involving minimal repair at failure with replacement after N failures. This cost function provides an appropriate criterion to determine the optimal replacement number N* for a system required for use over a finite time horizon. In an example, it is shwon that significant cost savings can be achieved using N* from the new finite time horizon model rather than the value predicted by the usual asymptotic model.