计算机工程与应用 ›› 2020, Vol. 56 ›› Issue (8): 270-278.DOI: 10.3778/j.issn.1002-8331.1911-0282

• 工程与应用 • 上一篇    

改进蚁群算法在AGV路径规划中的应用

胡春阳,姜平,周根荣   

  1. 南通大学 电气工程学院,江苏 南通 226019
  • 出版日期:2020-04-15 发布日期:2020-04-14

Application of Improved Ant Colony Optimization in AGV Path Planning

HU Chunyang, JIANG Ping, ZHOU Genrong   

  1. School of Electrical Engineering, Nantong University, Nantong, Jiangsu 226019, China
  • Online:2020-04-15 Published:2020-04-14

摘要:

针对传统蚁群算法在路径规划时,易陷入局部最优、前期路径有效性差等问题,对传统蚁群算法进行改进并应用到AGV(Automated Guided Vehicle)路径规划上。采用栅格地图建立小车工作空间模型,利用改进的头尾搜索机制,提高并加快了算法的全局搜索能力和前期收敛速度;引入奖惩因子与信息素最大最小阈值,对每代最优路径上的信息素进行奖励,最差路径上的进行惩罚,提高全局搜索能力;引入遗传算法变异因子,使算法跳出局部最优能力加强;采用遗传算法对改进的蚁群算法进行参数优化,减少参数对算法的影响。在VS2017和MATLAB软件平台上进行算法仿真。结果表明了该算法在避免局部最优和加快收敛速度方面有很大改进。

关键词: 路径规划, 改进蚁群算法, 奖惩因子, 参数优化

Abstract:

The traditional ant colony optimization in path planning is easy to fall into local optimization and poor early path effective solution, in order to solve this, an improved algorithm of the traditional ant colony optimization is proposed and applied to AGV(Automated Guided Vehicle) path planning. Firstly, the grid map is used to build the car workspace model, and the improved head and tail search mechanism is used to improve and accelerate the global search ability and early convergence speed of the algorithm. The maximum and minimum thresholds of reward and punishment factors and pheromones are introduced to reward the pheromones on the optimal path of each generation, and the punishment on the worst path is introduced to improve the global search ability. Genetic algorithm variation is introduced. The genetic algorithm is used to optimize the parameters of the improved ant colony optimization to reduce the influence of parameters on the algorithm. And the algorithm simulation is carried out on VS2017 and MATLAB software platform. The results show that the algorithm has great improvement in avoiding local optimization and accelerating convergence speed.

Key words: path planning, improved ant colony optimization, reward and punishment factor, parameter optimization