摘要
针对传统蚁群算法在复杂仓储环境中路径规划时不具备自主避障并且陷入局部最优解从而得不到最优路径的问题进行研究。文章通过构造MAKLINK无向网络图和优化蚁群算法中的启发函数,引入动态光滑因子和全局信息因子,达到控制AGV路径的长度和光滑程度,并构造避障因子使AGV路径在达到全局最优路径的同时拥有避开障碍的能力。通过研究,提出的改进算法可以实现自主避障,并且在最大迭代次数、路径距离和光滑程度方面优于传统蚁群算法。
Aiming at the problem that traditional ant colony algorithm does not have obstacle avoidance in the path planning in complex storage environment and falls into the localoptimal solution and can not get the optimal path. By constructing the MAKLINK undirected network graph and the heuristic function in the optimized ant colony algorithm, this paper introduces the dynamic smoothing factor and the global information factor to control the length and smoothness of the AGV path, and constructs the obstacle avoidance factor to make the AGV path reach the global maximum. The excellent path also has the ability to avoid obstacles. Through research,the improved algorithm proposed in this paper can achieve autonomous obstacle avoidance, and is superior to traditional ant colony algorithm in terms of maximum iteration numberand optimal path planning.
引文
[1]徐海军,潘迪.基于A*算法的无冲突路由多路AGV控制策略[J].工业控制计算机,2018,31(8):100-101.
[2]汤红杰,王鼎,皇攀凌,等.优化Dijkstra算法在工厂内物流AGV路径规划的研究[J].机械设计与制造,2018(5):117-120.
[3]项宏峰,曹少中,徐长波,等.基于神经网络的AGV智能车路径规划的仿真研究[J].北京印刷学院学报,2017,25(7):128-130.
[4]王娜.基于改进蚁群算法的多AGV作业调度研究[D].西安:陕西科技大学(硕士学位论文),2017.
[5]王松涛.基于优化的遗传算子改进蚁群算法AGV路径规划[J].系统解决方案,2016(10):47-49.
[6]苑光明,翟云飞,丁承君,等.基于改进遗传算法的AGV路径规划[J].北京联合大学学报,2018,32(1):65-69.
[7]李明富,张玉彦,马建华,等.基于变参数萤火虫算法和MAKLINK图的路径规划研究[J].机械科学与技术,2015,34(11):1728-1732.