Konferensartikel

This research studies the performance of an optimization technique for phase out optimization of technical systems. The optimization search is done by a genetic algorithm where good candidate samples can be drawn and evaluated in a fast manner using a novel concept denoted matrix simulation. The method is illustrated by applying it to a phase-out scenario of an aircraft fleet, where the optimal stop-maintenance strategy is determined for a set of repairable items (rotables). The main contribution of this paper is a parametric study of how the optimization search is expected to perform when varying the sampleset and problem size. In the parametric study the genetic algorithm result is compared against the real optimum which was found in a separate time-consuming brute force search, going through all possible solutions.

phase-out, maintenance optimization, genetic algorithm, matrix simulations