NEW CONTRIBUTIONS TO THE MODELLING AND OPTIMIZATION OF THE SELECTIVE MAINTENANCE PROBLEM IN COMPLEX SYSTEMS

Abstract

Selective maintenance is a strategy used in industrial and military environments where maintenance is performed between a sequence of missions. This thesis contributes to the SM literature by introducing and solving the joint selective maintenance and orienteering problem (JSMOP). The JSMOP can be applied to a wide range of systems that are geographically distributed and maintained by a crew of repair technicians. The JSMOP model will simultaneously decide the systems to visit and the maintenance actions to be performed on the visited systems to meet or exceed a target reliability. The multimission selective maintenance problem attempts to identify the maintenance actions to be performed across multiple missions. A second contribution is made by introducing a new solution method for the multimission SMP in the form of a hybrid column generation and genetic algorithm. Through numerical experiments the hybrid column generation and genetic algorithm is shown to outperform other metaheuristics.