Lookahead approximate dynamic programming for stochastic aircraft maintenance check scheduling optimization

Abstract

This paper proposes a lookahead approximate dynamic programming methodology for aircraft maintenance check scheduling, considering the uncertainty of aircraft daily utilization and maintenance check elapsed time. It adopts a dynamic programming framework, using a hybrid lookahead scheduling policy. The hybrid lookahead scheduling policy makes the one-step optimal decision for heavy aircraft maintenance based on deterministic forecasts and then determines the light maintenance according to stochastic forecasts. The objective is to minimize the total wasted utilization interval between maintenance checks while reducing the need for additional maintenance slots. By achieving this goal, one is also reducing the number of maintenance checks and increasing aircraft availability while respecting airworthiness regulations. We validate the proposed methodology using the fleet maintenance data from a major European airline. The descriptive statistics of several test runs show that, when compared with the current practice, the proposed methodology potentially reduces the number of A-checks by 1.9%, the number of C-checks by 9.8%, and the number of additional slots by 78.3% over four years.