Optimizing combination of aircraft maintenance tasks by adaptive genetic algorithm based on cluster search

It is significant to combine multiple tasks into an optimal work package in decision-making of aircraft maintenance to reduce cost, so a cost rate model of combinatorial maintenance is an urgent need. However, the optimal combination under various constraints not only involves numerical calculations but also is an NP-hard combinatorial problem. To solve the problem, an adaptive genetic algorithm based on cluster search, which is divided into two phases, is put forward. In the first phase, according to the density, all individuals can be homogeneously scattered over the whole solution space through crossover and mutation and better individuals are collected as candidate cluster centres. In the second phase, the search is confined to the neighbourhood of some selected possible solutions to accurately solve with cluster radius decreasing slowly, meanwhile all clusters continuously move to better regions until all the peaks in the question space is searched. This algorithm can efficiently solve the combination problem. Taking the optimization on decision-making of aircraft maintenance by the algorithm for an example, maintenance which combines multiple parts or tasks can significantly enhance economic benefit when the halt cost is rather high.