华南理工大学学报(自然科学版) ›› 2021, Vol. 49 ›› Issue (5): 102-110.doi: 10.12141/j.issn.1000-565X.200300

所属专题: 2021年机械工程

• 机械工程 • 上一篇    下一篇

多机器人任务分配调度的克隆选择算法

全燕鸣 何一明   

  1. 华南理工大学 机械与汽车工程学院,广东 广州 510640
  • 收稿日期:2020-06-08 修回日期:2020-11-26 出版日期:2021-05-25 发布日期:2021-04-30
  • 通信作者: 全燕鸣(1957-),女,教授,博士生导师,主要从事机器视觉、机器人自主导航与路径规划及多机器人任务分配与调度方法研究。 E-mail:meymquan@scut.edu.cn
  • 作者简介:全燕鸣(1957-),女,教授,博士生导师,主要从事机器视觉、机器人自主导航与路径规划及多机器人任务分配与调度方法研究。
  • 基金资助:
    广东省普通高校特色创新类项目(2019KTSCX003);广东省自然科学基金资助项目(2020A1515011503)

Research on Clonal Selection Algorithm for Multi-Robot Task Allocation and Scheduling

QUAN Yanming HE Yiming   

  1. School of Mechanical & Automotive Engineering, South China University of Technology, Guangzhou 510640, Guangdong, China
  • Received:2020-06-08 Revised:2020-11-26 Online:2021-05-25 Published:2021-04-30
  • Contact: 全燕鸣(1957-),女,教授,博士生导师,主要从事机器视觉、机器人自主导航与路径规划及多机器人任务分配与调度方法研究。 E-mail:meymquan@scut.edu.cn
  • About author:全燕鸣(1957-),女,教授,博士生导师,主要从事机器视觉、机器人自主导航与路径规划及多机器人任务分配与调度方法研究。
  • Supported by:
    Supported by the Characteristic Innovation Projects of Ordinary Colleges and Universities in Guangdong Province (2019KTSCX003) and the Natural Science Foundation of Guangdong Province(2020A1515011503)

摘要: 针对智能制造系统中同一时间段内多机器人执行多任务时,任务分配调度难以稳定控制及优化的问题,提出一种应用于多机器人任务分配调度的克隆选择算法。首先分析多机器人基于多任务的初始条件,以按生产过程完成任务为约束条件,以系统持续时间、单机器人最大消耗及多机器人总消耗为目标函数,构建多机器人任务分配调度优化的数学模型,使用克隆选择算法进行迭代求解,引入亲和度函数以动态改变克隆、变异、选择参数,从而提高计算效率;最后通过甘特图进行分析并验证此调度方法。结果表明此方法稳定可行;实验证明在多任务条件下,克隆选择算法收敛速度快、计算精度高,具有同时优化多项指标的优点。

关键词: 智能制造系统, 多机器人, 多任务分配, 调度, 克隆选择算法

Abstract: A clonal selection algorithm (CSA) for multi-robot collaborative scheduling was proposed to solve  the difficulties in controlling and optimizing the task assignment and scheduling of multi-robot in the same period in intelligent manufacturing system. Firstly, the initial conditions of multi-robot task were analyzed. Then taking the task completion as the constraint condition and taking task flow time, single robot maximum cost and multi-robot total cost as the objective function, a mathematical model of multi-robot task allocation and scheduling optimization was constructed. The affinity function was introduced to dynamically change the parameters of clone, mutation and selection, so as to improve the computational efficiency. Finally, the scheduling method was verified by Gantt chart analysis. The results show that the proposed method is stable and feasible. Meanwhile, the experiments show that the clonal selection algorithm has the superiority of fast convergence speed and high calculation accuracy, and can optimize various indicators under complex multi-task conditions.

Key words: intelligent manufacturing system, multi-robot, multi-task allocation, scheduling, clonal selection algorithm

中图分类号: