Sequential nonparametric age replacement policies (Q1069254)

Consider an age replacement policy, by which a stochastically failing unit is replaced at failure or after being in service for t units of time, whichever comes first. The problem is to estimate sequentially the optimal replacement time \(\phi^*\) which minimizes the expected long-run average cost. A recursive estimator \(\phi^*_ n\) of stochastic approximation (SA) type is proposed. Using theorems of \textit{H. Robbins} and \textit{D. Siegmund} [Optimizing Meth. Statist., Proc. Sympos. Ohio State Univ. 1971, 233-257 (1971; Zbl 0286.60025)], and \textit{V. Fabian} [Ann. Math. Stat. 39, 1327-1332 (1968; Zbl 0176.484)], the authors prove that the proposed estimator achieves the best long-run cost, and derive the rate of convergence of the optimal replacement time \(\phi^*\). A Monte Carlo study is presented to demonstrate the usefulness of the proposed procedure. Comparisons to a sequential procedure by \textit{J. A. Bather}, [Bull. Inst. Int. Stat. 47, 253-266 (1977)] are briefly discussed.