A Decision Making Model for Replacing Non-Repairable Systems

The classical model for finding optimal preventive replacement policies is revisited. In minimizing expected cost per unit time in a cycle, we first introduce discount factors since the present value of the replacement cost depends on the time when replacement actually occurs. Then certainty equivalent is used to take risk into account arising from the randomness of the cost per unit time, since the time of failure is uncertain. A simple numerical example with Weibull variable illustrates the methodology.