This paper addresses multiple unmanned aerial vehicle (UAV) task allocation in post-disaster rescue scenarios by proposing a confidence-driven multi-objective particle swarm optimization algorithm (CD-MOPSO). The method incorporates constraints on task time, UAV payload, and flight range, formulating a mathematical model to minimize task failures and energy consumption. Innovations include a novel particle decoding method, an enhanced particle mutation strategy, refined global best selection, and improved external archive maintenance to boost archive quality. A confidence-driven velocity update strategy based on an emotion contagion mechanism, dynamically balancing exploration and exploitation based on neighboring particle states and fitness changes. Extensive simulation, ablation studies, and comparative experiments demonstrate the algorithm's effectiveness. Compared with conventional multi-objective PSO(MOPSO), improved multiple objective particle and genetic algorithm, non-dominated sorting genetic Algorithm II, and consensus-based bundle algorithm, the proposed method achieves HV metric improvements of 7%, 13.8%, 218%, and 73.2% across four scenarios, respectively. Further, a test platform based on PIXHAWK 2.4.8 flight controller was established, and the real-world experiment verified its effectiveness. This work offers an effective solution for UAV task allocation in disaster response, enhancing rescue operations.
Sun et al. (Sun,) studied this question.