平衡探索与利用的广义鸽群优化算法

In order to balance the exploration and exploitation capabilities of the pigeon-inspired optimization algorithm, a generalized pigeon-inspired optimization algorithm is proposed. The traditional pigeon-inspired optimization algorithm contains two optimization operators, namely the map and compass operator and the landmark operator. These two operators are executed sequentially and only one round of iteration is performed in one run of the algorithm. In the generalized pigeon-inspired optimization algorithm, the algorithm search is divided into multiple stages, and each stage executes the two operators respectively. In one run of the algorithm, the two operators are executed for multiple rounds. The map and compass operator focuses on the exploration ability of the algorithm, while the landmark operator focuses on the exploitation ability of the algorithm. The improved algorithm only changes the execution order of the two operators and does not require additional function value calculations. In addition, the generalized pigeon-inspired optimization algorithm extends the solution set structure and operator parameter settings, which is of great benefit to improving the search quality of the algorithm. Simulation comparison experiments were carried out on 11 single-objective test functions and 8 multi-modal optimization test functions. The results show that the generalized pigeon-inspired optimization algorithm improves the search efficiency of the pigeon-inspired optimization algorithm and improves the search results of the algorithm.

为了平衡鸽群优化算法的探索与利用能力, 提出了一种广义鸽群优化算法. 传统的鸽群优化算法包含两种优化算子, 分别为地图与指南针算子与地标算子. 这两种算子依次执行, 在一次算法运行中, 仅执行一轮迭代. 在广义鸽群优化算法中, 将算法搜索分为多个阶段, 每个阶段分别执行两种算子. 在算法的一次运行中, 两种算子执行多轮. 地图与指南针算子侧重于算法的探索能力, 而地标算子侧重于算法的利用能力. 改进算法仅改变了两种算子的执行顺序, 无需增加额外的函数值计算. 此外, 广义鸽群优化算法扩展了解集合结构和算子参数设置, 这对于提高算法的搜索质量大有裨益. 在 11 个单目标测试函数和 8 个多模态优化测试函数上进行了仿真对比试验, 结果表明广义鸽群优化算法提高了鸽群优化算法的搜索效率, 改进了算法的搜索结果.