An improved multi-objective restart variable neighborhood search algorithm for aircraft sequencing problem with complex interdependent runways

Abstract

The Aircraft Sequencing Problem (ASP) is a problem that seeks to minimize adverse impacts such as flight delays by adjusting flight sequencing schemes. With the increase in aviation demand, ASP has been garnered more attention in recent years. However, existing ASPs typically assume that the relationships between runways are either entirely independent or entirely interdependent, without considering the complex scenarios where runways have both dependent and independent relationships. Moreover, existing ASP research focusing on multi-objective aspects is relatively scarce, while it is essential to consider multifaceted needs in reality. Therefore, we have established a multi-objective mathematical model for the Aircraft Sequencing Problem with Complex Interdependent Runways (ASP-CIR), aiming to minimize the total delays and makespan. Different from the common lack of problem-specific structural design in existing ASP algorithms, or the need to improve structural strategies, we propose an Improved Multi-Objective Restart Variable Neighborhood Search (IMORVNS) algorithm based on the characteristic of the studied ASP-CIR. In the IMORVNS, we have designed several problem-specific structures guided by constraints, such as initialization and neighborhood search. We tested our new method on 30 different problem instances of three sizes and found it worked better than other recent algorithms. Moreover, a sensitivity analysis on the scenario featuring both interdependent and independent runway relationships was conducted, reaffirming the importance of considering the ASP-CIR scenario.