Abstract: The rapid development of modern defense technology decreases the USVs’ attacking effect greatly, autonomous formation cluster attack technique of USVs has become one of the key technologies of future naval warfare, mission planning among USVs is the key for them to complete tasks smoothly and efficiently. It regards the cluster attack mission planning problem as multi-constrained task allocation process, builds mission planning model. A PSO optimization algorithm based on distributed auction is proposed, this algorithm improves particle initialization and optimization process combined with distributed auction mechanism to make partial meet mission constrain condition and maintain diversity, this will avoid PSO optimization falling into a local optimum. Simulation result indicates that the program achieved with distributed auction mechanism particle swarm optimization could fully meet the requirements of USVs’ cluster attack missions, and shows better convergence compared with traditional particle swarm optimization and other swarm intelligence algorithms.

Key words: multi-Unmanned Surface Vehicle（USV）, mission planning, distributed auction mechanism, Particle Swarm Optimization（PSO）

摘要： 现代防御技术的迅速发展使得水面舰艇的攻击效果大大下降，水面无人舰艇自主编队集群攻击技术已经成为未来海战的关键技术之一，多水面无人舰艇之间的任务规划是保证无人舰艇顺利、高效完成任务的关键。将水面无人舰艇集群攻击任务规划问题看成是多约束的任务分配过程，建立任务规划模型，提出了基于分布式拍卖机制的粒子群优化算法，该算法结合分布式拍卖机制对粒子群优化算法的粒子初始化和寻优过程进行改进，使得粒子既符合任务的约束条件，又保持了多样性，避免粒子在寻优过程中陷入局部最优。仿真结果表明应用分布式拍卖机制粒子群优化算法得到的方案不仅完全满足水面无人舰艇集群攻击任务的要求，而且比传统粒子群优化算法和其他群体智能算法具有更好的收敛性。

关键词: 多水面无人舰艇, 任务规划, 分布式拍卖机制, 粒子群优化