针对有人/无人协同背景下多UUV任务分配存在的任务负载不均衡、动态适应性不强等问题，基于群体智能提出一种动静结合的启发式任务分配方法，以UUV为主体设置主循环，循环往复地为各UUV分配满足优化目标和分配约束的任务，直至资源耗尽或任务完成为止。利用阈值响应机制，保证基本的任务分配质量；引入轮盘赌随机策略，提升任务分配方案的全局最优性；设置每轮分配任务数上限，改善任务分配方案的均衡性；提出任务重置和任务接续2种策略，赋予算法动态重分配能力。基于典型想定进行仿真实验，结果表明，提出的任务分配方法既能进行全局静态任务分配，又能应对UUV故障、新增任务等动态应用场景。
Aiming at the problems of unbalanced task load and weak dynamic adaptability in multi-uuv task allocation under the background of manned/unmanned cooperation, a dynamic static heuristic task allocation method is proposed based on swarm intelligence. The main cycle is set with UUV as the main body, and the tasks that meet the optimization objectives and constraints are allocated to each UUV repeatedly until the resources are exhausted or the tasks are completed. The threshold response mechanism is used to ensure the basic task allocation quality. The random roulette strategy is introduced to improve the global optimality and the upper limit of tasks assigned is set to improve the balance of task allocation scheme. Besides, two strategies, task reset and task continuation, are proposed to realize dynamic redistribution capability. Simulation experiments based on typical scenarios show that the proposed task allocation method can not only carry out global static task allocation, but also deal with dynamic application scenarios such as UUV failure and new tasks.
2023,45(1): 94-100 收稿日期：2021-12-27
DOI：10.3404/j.issn.1672-7649.2023.01.017
分类号：E919
基金项目：中国博士后科学基金项目资助（2021M693939）
作者简介：范学满(1989-),男,博士,助理研究员,研究方向为智能辅助决策、无人集群任务规划
