Periodic replacement strategies: optimality conditions and numerical performance comparisons

摘要

In maintenance engineering, age replacement policy (ARP) and block replacement policy (BRP) are the most popular basic strategies. They have been intensively studied and compared using different performance measures. Several of these comparisons are stochastics on the basis of the renewal theory, and a few of them are of economic benefit. This paper presents a comparative study for analysing ARP and BRP models using the expected costs function as the principal criterion. To provide this comparison, we propose a numerical approach allowing to combine cost/distribution for the determination of the optimal strategy. For that, we resume the main analytical results and prove that a finite solution exists if the failure rate increases. Results clearly show that both strategies are very close, which intuitively confirm the statement of Barlow and Proschan's theorem. Based on the computational results, we show that the ultimate decision to select the best strategy is conditioned by the choice of the distribution function, the value of its parameters and that the periodic replacement unit cost must be much lower than the replacement unit cst at failure.