Optimal Replacement under Differing Amounts of Information

摘要: Consider a coherent system (see e.g. Barlow and Proschan [2]) consisting of n components with stochastic lifetimes. Upon failure the has to be replaced at cost c > 0 an additional penalty k 0. The may before only c. At any time states fixed subset B set C components, state age in service are known. A replacement policy is required that uses this information minimizes long-run average per unit time. Obviously we have usual model if empty set, none observed. This been examined extensively, by [1]. Beckmann [3] shows how use dynamic programming for finding optimal rule. He also studies minimization total discounted cost.