面向多类型资源需求的机器人任务分配方法  

A robot task allocation method for multi-type resource requirements

作　　者：何杏宇 吴西博 杨桂松[1] 贾明权[3] HE Xingyu;WU Xibo;YANG Guisong;JIA Mingquan(School of Optical-Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China;College of Communication and Art Design,University of Shanghai for Science and Technology,Shanghai 200093,China;Southwest China Institute of Electronic Technology,Chengdu 610036,China)

机构地区：[1]上海理工大学光电信息与计算机工程学院,上海200093 [2]上海理工大学出版印刷与艺术设计学院,上海200093 [3]中国西南电子技术研究所,成都610036

出　　处：《智能计算机与应用》2024年第6期11-19,共9页Intelligent Computer and Applications

基　　金：国家自然科学基金(61802257,61602305);上海市自然科学基金(18ZR1426000,19ZR1477600);敏捷智能计算四川省重点实验室开放式基金资助项目。

摘　　要：针对多机器人系统中任务的多类型资源需求问题,单个机器人往往因能力有限而无法胜任,需要多个机器人来协同完成。为此,提出了一种面向多类型资源需求的任务分配优化模型,在单个机器人携带资源有限的情况下,为任务分配合适的机器人协同小组。该模型综合考虑了影响任务协同成本的时间、移动距离、资源利用率等因素,设计了任务的协同成本函数,并提出基于纳什均衡策略的任务分配方法。该方法首先通过遗传算法为更多的任务分配符合其资源需求的初始机器人小组,然后利用博弈论中的纳什均衡思想为初始机器人小组寻找最优平衡解,最终得到满足纳什均衡解的机器人小组,从而降低任务的协同成本。实验结果表明,本文所提方法能有效提高任务完成率和降低任务的平均协同成本。In view of the multi-type resource demand of tasks in multi-robot systems,individual robots are often unable to perform the task by their limited abilities,so that multiple robots are required to complete cooperatively.To address the problem,this paper proposes a task allocation optimization model for multiple resource requirements to assign the task to a suitable robot cooperative group with individual robot limited resources.This model designs the collaboration cost function of the task which comprehensively considers time,travel distance,resource utilization and other factors affecting task collaboration cost.And the model proposes a task assignment method based on the Nash equilibrium strategy.Firstly,this method assigns more tasks to initial robot groups that met their resource needs by Genetic Algorithm.Then,the optimal equilibrium solution is achieved for the initial groups by Nash equilibrium thought in Game Theory.Finally,this method obtains the robot groups that meets the Nash equilibrium solution,so as to reduce the collaborative cost of tasks.The experimental results show that the proposed method can effectively improve the task completion rate and reduce the average collaborative cost of the task.

关 键 词：多机器人系统 任务分配 协同小组 遗传算法 纳什均衡解 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]