Multi-objective hub location for urban air mobility via self-adaptive evolutionary algorithm

As urbanization increases and city populations grow, traditional metropolitan logistics face challenges in managing the rising volume of goods efficiently. Urban air mobility (UAM) offers a promising solution for urban goods delivery by enabling fast and direct aerial transport between key hubs, forming the basis for urban air logistics (UAL). However, most existing studies on UAL hub location primarily focus on solely optimizing economic factors such as cost and efficiency, often overlooking critical safety concerns associated with third-party risks due to UAM operations. To fill this gap, we develop a novel multi-objective mixed-integer programming model that integrates a comprehensive urban air delivery risk assessment with the hub location problem. Our model simultaneously considers economic costs and third-party safety risks as bi-objectives, providing trade-off solutions between economic factors and safety concerns for decision makers. Furthermore, to effectively solve the model, we propose an improved non-dominated sorting genetic algorithm with adaptive operator selection (INSGA-AOS). This algorithm incorporates a bi-level encoding scheme, nine problem-specific search operators for crossover and mutation, and an adaptive operator selection mechanism based on multi-armed bandits (MABs) to dynamically identify the optimal combinations of crossover and mutation operators. Experimental results on instances generated from a benchmark dataset demonstrate that our proposed algorithm significantly outperforms existing methods in generating high-quality Pareto front. Further analysis verifies the effectiveness of different components within the proposed method.