A Multi-Objective Approach to Optimize a Periodic Maintenance Policy

The present paper proposes a multi-objective approach to find out an optimal periodic maintenance policy for a repairable and stochastically deteriorating multi-component system over a finite time horizon. The tackled problem concerns the determination of the system elements to replace at each scheduled and periodical system inspection by ensuring the simultaneous minimization of both the expected total maintenance cost and the expected global system unavailability time. It is assumed that in the case of system elements failure they are instantaneously detected and repaired by means of minimal repair actions in order to rapidly restore the system. A non-linear integer mathematical programming model is developed to solve the treated multi-objective problem whereas the Pareto optimal frontier is described by the Lexicographic Goal Programming and the ε-constraint methods. To explain the whole procedure a case study is solved and the related considerations are given.