摘要
针对动态仓储环境下多机器人运动过程中出现的拥塞死锁问题,利用路径长度、转弯数、路径惩罚函数建立小车单任务耗时模型。模型引入阻塞惩罚函数,移除可能发生阻塞的路径增加罚值。同时针对传统遗传算法路径规划操作过程中路径交叉变异导致路径中断不可用的情况,设计重复点交叉算子,在变异操作后检查路径合法性,使算法都是在可行的解空间上进行搜索。仿真实验表明,算法能指导机器人获得动态环境下的最优路径,同时算法收敛速度大大提高。
In order to solve the problem of congestion deadlock during the movement of multiple robots in a dynamic storage environment,the path length,the number of turns,and the path penalty function are used to build a robot single task time-consuming model.The model introduced a blocking penalty function to remove the path that may be blocked and increase the penalty value.At the same time,in response to the path unavailability caused by path cross mutation during the path planning operation of traditional genetic algorithm,a repeat point crossover operator was designed to check the legitimacy of the path after the mutation operation,so that the algorithm searched on a feasible solution space.Simulation experiments show that the algorithm can guide the robot to obtain the optimal path in a dynamic environment,and the convergence speed of the algorithm is greatly improved.
作者
杨世团
于宝成
吴云韬
Yang Shituan;Yu Baocheng;Wu Yuntao(Hubei Key Laboratory of Intelligent Robot,Wuhan 430205,Hubei,China;School of Computer Science&Engineering,Wuhan Institute of Technology,Wuhan 430205,Hubei,China)
出处
《计算机应用与软件》
北大核心
2022年第3期56-62,101,共8页
Computer Applications and Software
基金
国家自然科学基金面上项目(61771353)。
关键词
动态仓储环境
单任务耗时模型
阻塞惩罚函数
遗传算法
Dynamic storage environment
Single task time consuming model
Congestion deadlock
Genetic algorithm
