Local Search Heuristic for the Optimisation of Flight Connections

Kiwi.com proposed a real-world NP-hard optimisation problem with a focus on air travelling services, determining the cheapest connection between specific areas. Despite some similarities with the classical TSP problem, more complexity is involved that makes the problem unique. It is Time-dependent, Asymmetric and involves areas that contain sets of cities from which exactly one is visited. In addition to this, infeasibility adds more complexity to the problem since there are no flights available between specific points in the network for certain days. While solving such computationally difficult problems, exact methods often fail, particularly when the problem instance size increases; Then alternative approaches, such as heuristics, are preferred in problem solving. In this study, we present an effective local search method for solving Kiwi.com problem. The empirical results show the success of the approach, which embeds four simple operators, on most of the released instances.