Abstract
In this paper, an adaptive replacement policy for a finite time span is considered, assuming that the failure distribution is a Weibull one with unknown scale parameter. This policy is characterized by the fact that after each replacement a new planned replacement interval is selected so as to minimize the expected expenditure during the remaining time, taking in account the previous information and new data obtained, after each replacement, about the unknown parameter. Using some properties derived from this policy, the authors found the optimal replacement interval numerically.
