Performance evaluation of seven optimization models of age replacement policy

Several optimization models of the classic age replacement policy have been developed based on different optimality criteria in the literature. This raises the issue how to choose an appropriate model among these models for a specific application. This paper addresses this issue by carrying out a comparative study for the performances of these models and a new risk-sensitive cost model. The performance of the solution of an optimization model is evaluated from two aspects: cost (or availability) and reliability. Two benchmarking points that correspond to these two aspects are defined; and three performance measures that correspond to cost, reliability and overall performance, respectively, are developed for evaluating the performances of the models. Through quantitatively comparing the performances of the models' solutions associated with the Weibull, gamma and lognormal distributions, the models with outstanding performances are identified. The applicable situations of the outstanding models are also discussed. The results are useful for both maintenance researchers and practitioners.