Review of a Markov decision algorithm for optimal inspections and revisions in a maintenance system with partial information

In a previous published paper Tijms and van der Duyn Schouten proposed a special-purpose policy iteration algorithm for optimal inspections and revisions in a maintenance system with partial information. In this paper the authors point out some shortcomings of this algorithm. They demonstrate that their algorithm, which operates on a class of control-limit rules, does not always lead to an optimal policy within that class, and present appropriate improvements. In doing so, the authors analyse some specific aspects of the policy iteration algorithm.