2019 Volume 70 Issue 3 Pages 147-156
In recent years, companies have been challenged with the growing demand to design more effective airline schedules both domestically and internationally. Typically, the airline scheduling problem has been sequentially decomposed into the following four stages: (1) Flight Scheduling Problem (FSP), (2) Fleet Assignment Problem (FAP), (3) Aircraft Maintenance Routing Problem (AMRP), and (4) Crew Scheduling Problem (CSP). Unfortunately, in most of the previous research, these problems have been solved independently for reasons of computational tractability. Since the decisions from one stage impose upon the decision-making process in subsequent stages, this approach may result in sub-optimal solutions, and there is growing consensus of the necessity for an integrated view. In this study, we propose a model that integrates departure re-timing, fleet assignment, and aircraft maintenance routing decision-making. Allowing flexibility in the departure times of scheduled flight legs determined in a FSP can increase connection opportunities in the FAP and CSP, thus saving fleet assignment costs. There are a few papers that integrate these three decisions, but our model is differentiated by the following two aspects. First, we consider the detailed schedule of each fleet, while previous papers only considered the number and type of fleet assigned to each flight leg. Second, in our model, we impose the constraint that maintenance must be conducted at pre-specified airports that have maintenance capability. By doing so, the problem becomes much more complicated and an exact algorithm cannot be applied. We propose an ant colony optimization-based algorithm that exploit the special structure of the problem. A numerical example is utilized to illustrate the model, and a sensitivity analysis of the re-timing parameters is conducted to demonstrate the effectiveness of the model proposed.
