论文部分内容阅读
同时考虑蚁群算法的所有运行参数,提出一种基于图知识迁移的蚁群算法参数选择方法.首先,将包含知识(蚁群算法的运行参数)的源任务映射到一个高维的迁移空间,并通过迁移权值连接不同的源任务,构造一个模型迁移图;然后,扩展模型迁移图使其包含目标任务,并利用图论的知识学习迁移函数;最后,通过最小二乘法自主地给目标任务分配一个优化的运行参数组合.机器人路径规划问题的仿真结果验证了该方法的智能性、快速性与合理性.
At the same time, considering all operating parameters of ant colony algorithm, a parameter selection method of ant colony algorithm based on graph knowledge transfer is proposed.First, the source task containing knowledge (ant colony algorithm operating parameters) is mapped to a high-dimensional migration space, And construct a model migration diagram by connecting different source tasks by moving the weights; then, extend the model migration map to include the target tasks and use the knowledge of graph theory to learn the transfer function; finally, autonomously assign the target tasks by the least square method An optimal combination of operating parameters is assigned.The simulation results of the robot path planning problem verify the intelligence, rapidity and rationality of this method.