Models for aircraft maintenance routing problem with consideration of remaining time and robustness. (November 2019)

• Record Type: Journal Article
• Authors: Cui, Ruyu; Dong, Xingye; Lin, Youfang
• Is Part Of: Computers & industrial engineering. Volume 137(2019)
• Publication Date: 2019-11
• Subjects: Aircraft maintenance routing -- Integer linear programming -- Mixed integer linear programming -- Robustness -- Variable neighborhood search; Engineering -- Data processing -- Periodicals; Industrial engineering -- Periodicals; 620.00285
• DOI: 10.1016/j.cie.2019.106045
• Languages: English
• ISSNs: 0360-8352

Abstract

Highlights: Three MIP models are established aiming at the AMRP. A heuristic of the variable neighborhood search (VNS) algorithm is improved. Compared with CPLEX and the VNS performs better. Abstract: Aircraft maintenance routing problem is especially important with respect to timing and costs for an airline company. This work establishes an integer linear programming (ILP) model (Model I) for aircraft maintenance routing problem with the objective to minimize the number of aircrafts and total remaining flying time. Then, we extend the Model I to Model II and Model III, respectively. In the Model II, we consider the robustness based on aircraft delay probability, and add the robust constrains to the model aiming at reducing total aircraft delay costs. For the Model III, we take into account the fleet type of the aircrafts, whose objective is to minimize the total flying cost. In terms of solution method, we improve a heuristic of the variable neighborhood search (VNS) algorithm to solve the Model I and Model II which can generate a suboptimal solution quickly in reasonable time. In order to verify the effectiveness of the heuristic algorithm, the experimental results are compared with the commercial software solver CPLEX. The experimental results on real world flight network show that the performance of the heuristic algorithm is much better than CPLEX in terms of the computation efficiency for large scale test instance. Moreover, the heuristic algorithm also can obtain high quality solutions.