A Braess’s Paradox inspired method for enhancing the robustness of air traffic networks

Abstract

Air traffic networks (ATNs) play an important role in air transport. It is of practical application values to improve the robustness of ATNs. Here we propose a counter-intuitive idea with the inspiration comes from the Braess’s Paradox phenomenon. To be specific, we propose to delete edges from an ATN to improve its corresponding robustness. To achieve this goal, we formulate a bi-objective optimization problem which aims to maximize the robustness of the focal ATN as well as to minimize the number of edges to be removed. In order to address the developed optimization model, we introduce the nondominated sorting genetic algorithm (NSGA-II) and modify its algorithm operators to make it fit for the established model. To check if the research idea proposed works or not, we conduct experiments on nine real-world ATNs. In the experiments, NSGAII has been compared against its successor–NSGA-III, and another state-of-the-art optimization algorithm named MODPSO. Experiments indicate that NSGA-II performs better than the rest two algorithms on the tested ATNs. For the tested ATNs, three networks have their robustness improved by 100% by removing less than six edges while the remaining six get an improvement of around 10%. This work provides aviation decision makers with a new perspective on ATNs design and management.