A Multi-objective Memetic Algorithm for Network-Wide Flights Planning Optimization

The ever-growing air traffic flow has brought about great challenges to balance the airspace congestion and air traffic demands. This fact sparks numerous studies on Network-wide Flights Planning Optimization (NFPO) which aims to reconcile the contradiction between flight delay cost and airspace congestion by optimizing the pre-strategic flight plans from a network-wide point of view. In consideration of bi-objective and large-scale characteristics of the NFPO problem, this paper proposes a multi-objective Memetic Algorithm with Rerouting Meme (MARM) that incorporates an evolutionary global search framework with a problem-specific meme ulocal search operator. With the idea of heuristically reducing the interactions among flight trajectories to decongest the airspace, the Trajectories Correlation (TC) is defined as key network-wide knowledge and is applied to design the critical Rerouting Meme (RM). Additionally, to balance the ability of exploitation and exploration, the idea of simulated heating configuration setting is adopt for RM to integrate with the global search. Extensive empirical studies conducted on real large-scale traffic data of China air traffic network and flight plans support that MARM is beneficial to the NFPO problem via showing the improvement on effectiveness.