Optimisation of military aircraft engine maintenance subject to engine part shortages using asynchronous metamodel-assisted particle swarm optimisation and Monte-Carlo simulations

Abstract

ABSTRACTComprising the coordination of various maintenance levels, flight maintenance planning (FMP) constitutes an appealing and hard-to-solve optimisation problem. In this article, a modified version of the typical FMP problem is studied for a typical-size military fleet, focusing on engine maintenance: a disruption of the depot level maintenance (DLM) facilities is considered, causing long-term shortage of serviceable engine parts. In this context, an optimisation strategy is proposed to schedule engine and part exploitation so that engine availability be retained until the recovery of the DLM supply lines, and flight and maintenance requirements be fulfilled to the maximum possible extent. To incorporate the effects of stochastic events, model their interaction with the capacity of the available maintenance resources and assess the outcome of the proposed plans, Monte-Carlo simulations are employed. A computationally efficient variant of the asynchronous particle swarm optimisation method is introduced to find the optimal solutions, combining parallelisation of the evaluation calls with a metamodel-based pre-evaluation strategy to reduce the optimisation turnaround time. Tested on a full-scale application, the proposed method is shown to be capable of producing robust FMP solutions and, also, quantifying the maintenance chain's operating status to aid in long-term planning.