Comprehensive Improved Cooperative Particle Swarm Optimizer for USVs Task Allocation for Detection Missions

Task allocation is the key technology and challenge for unmanned surface vehicles’ (USVs) detect mission. Aiming to improve the rapidity and optimality of task allocation and path planning of USVs formation, this paper proposes comprehensive improved cooperative particle swarm optimizer (CICPSO) task allocation algorithm for USVs formation based on particle swarm optimization (PSO). Firstly, a chaos-based Logistic map is adopted to improve the initial solution of particle swarm, and a practical initial solution generation strategy is designed based on the specific detect mission. Then the swarm population is divided into two cooperative subpopulations that focused on exploration and exploitation respectively. The two subpopulations use different position updating strategies. A chaotic mapping based regeneration strategy is added to exploration-subpopulations to retain the diversity of the whole population. The full social learning strategy of the exploitation-subpopulation ensures rapid convergence and improves efficiency of exploitation. Besides, this algorithm integrates the comprehensive learning (CL) strategy to ensure its superiority in solving high-dimensional problems. Experiments on the basic numerical function validate the effectiveness and efficiency of CICPSO. Finally, the algorithm is simulated in the task allocation environment of USVs formation, and the results show that the algorithm has certain practicability.