多无人艇协同目标分配算法研究

Research on Cooperative Target Assignment Algorithm for Multiple USVs

摘要:

未来战争中很多作战任务需要由多艘无人艇相互配合才能完成，多无人艇协同目标分配是无人艇自主协同控制研究的关键技术之一。为了解决多参数、多约束条件下的目标分配问题，改进贝叶斯优化算法中网络构造方式及需要存储大量数据的不足，提出了基于决策图贝叶斯优化算法(Bayesian Optimization Algorithm with Decision Graphs，DBOA)的多无人艇协同目标分配方法。根据无人艇的消耗、目标价值的毁伤和执行任务预计耗费时间 3 个决策变量，并结合约束条件构建了多无人艇协同目标分配数学模型。仿真实例表明，DBOA 算法收敛速度快，能够达到全局最优解，基于 DBOA 的协同目标分配方法具有良好的时间效率和分配效果。

Abstract:

In the future war，many combat missions need to be completed by the cooperation of multi-USV. Cooperative target assignment of multi-USV is one of the key technologies for research on autonomous cooperative control of USV. In order to solve the target assignment problem with multi parameters and multi constraints，and to improve the network construction method and the situation of lacking large amount of data storage in Bayesian optimization algorithm，a multi-USV cooperative target assignment method based on Bayesian optimization algorithm with decision graphs(DBOA)is proposed. According to USV attrition，target value damage and task expending time，a cooperative target assignment model for multi-USV is constructed. Simulation results show that the DBOA has fast convergence speed and can achieve global optimal solution. The cooperative target assignment method based on DBOA has good time efficiency and assignment effect.