摘要
为了提高花授粉算法的寻优精度和性能稳定性,提出了融入改进策略的精英协作引导花授粉算法。分析了花授粉算法的工作原理及缺陷,给出了对立点初始化方法提高初始种群质,使用精英协作引导全局搜索策略提高算法搜索效率和质量,建立动态转换概率模型用于平衡全局搜索与局部搜索,综合以上改进措施提出了精英协作引导花授粉算法。使用单模态标准函数和多模态标准函数对算法性能进行测试,每一项改进措施都能够提高算法的搜索精度和寻优稳定性,且改进措施之间不存在抵消性耦合。将精英协作引导花授粉算法应用于机器人路径规划,与花授粉算法相比,路径平均长度减少了4.01%,标准差减少了一个数量级以上,表明改进算法在路径规划质量和稳定性方面均优于传统花授粉算法。
In order to improve optimization accuracy and stability of flower pollination algorithm,elite collaboration guided flower pollination algorithm introducing improving strategy is proposed.Working principle and shortcomings of flower pollination algorithm are analyzed.Opposing point initialization method is given to improve initial population quantity.Elite collaboration guided strategy is used to improve searching efficiency and quantity.Dynamic adaptive transition probability model is built to balance global searching and local searching.Combining the improved measures above,elite collaboration guided flower pollination algorithm is put forward.Using single mode standard function and multi-mode standard function to test algorithm property,every improved measure can improve algorithm property,and it does not exist canceling coupling between different improving measures.Using elite collaboration guided flower pollination algorithm to plan robot path,compared with flower pollination algorithm,average path length decreases by 4.01%,and standard deviation decreases by more than one order magnitude,which means path quality and stability of improved algorithm are optimal to traditional flower pollination algorithm.
作者
朱旭东
ZHU Xu-dong(Wuxi Institute of Art and Technology,Jiangsu Yixing 214200,China)
出处
《机械设计与制造》
北大核心
2021年第10期241-245,共5页
Machinery Design & Manufacture
基金
江苏省教育厅“青蓝工程”资助项目(无编号)。
关键词
精英协作引导
对立点初始化
动态自适应转换概率
花授粉算法
机器人路径规划
Elite Collaboration Guided
Opposing Point initialization
Dynamic Adaptive Transition Probability
Flower Pollination Algorithm
Robot Path Planning
