Improving classical optimal age-replacement policies for degrading items

Abstract

We consider items with observable at inspections degradation. Following an inspection, a decision is made whether to replace an item or to continue operation. When degradation is relatively small, it is cost-beneficial to continue operation and postpone the preventive maintenance. The innovative, probabilistically justified replacement policy defines the postponement in this case. It is based on comparison of the observed degradation with the specially defined reference values. Degradation is modeled by the increasing stochastic process, specifically by the Poisson counting process and by the homogeneous gamma process. Detailed illustrative numerical examples describing the main steps of the developed original methodology are provided. They also show that the proposed policy can result in a significant cost reduction and increase in the replacement times.