Multi-Objective Immune Algorithm for Multi-UAV Patrol Task Allocation

To solve the cooperative patrol task allocation problem for multiple unmanned aerial vehicles (UAVs), this paper first establishes a multi-objective optimization model for patrol task allocation. The model considers three key factors: the importance level of each patrol node, the waiting time of UAVs, and the endurance constraints of UAVs. The objectives are to minimize the total patrol time, total patrol distance, and idle time of each patrol node. Considering the model's complexity, we design specialized encoding and decoding methods and an invalid chromosome repair mechanism. In addition, cubic chaotic mapping is introduced into the encoding generation process to enhance optimization performance. Subsequently, this paper proposes a Multi-Population Multi-Objective Immune-Genetic Algorithm (MPMOIA-GA) that integrates the clone selection operator from immune optimization algorithms with crossover and mutation operators from genetic algorithms. During iterations, the algorithm implements distinct evolutionary operations for replicated elite populations and ordinary populations, thereby enhancing both global and local search capabilities. Additionally, random chromosome generation is introduced to refresh populations and improve diversity. Finally, the experimental results show that the proposed MPMOIA-GA outperforms NSGA-II in terms of solution effectiveness, and can effectively enhance the solving efficiency.