Optimal, opportunistic maintenance policy using genetic algorithms,2: analysis

Investigates the use of a genetic‐algorithm program for analysing the optimal opportunity‐based maintenance problem for real‐sized systems. Analyses the performance of the genetic operators with a generation replacement genetic algorithm, using a hypothetical system consisting of 50 maintenance‐significant parts. Due to the size of the problem and excessive running time, finds that the steady‐state genetic algorithm gives the best compromise between solution quality and running time and was subsequently implemented for this problem. Pays special attention to the sensitivity of solutions to the maximum number of maintenance groups considered by the genetic algorithm. Finds that better solutions were identified for larger numbers of groups but increasing complexity costs more in terms of the computer time required. Also concludes that the improvement in the objective function value decreases with the increase in the number of maintenance groups.