关键词: 花授粉算法, 混沌映射, 动态全局搜索, Cauchy变异

Abstract: Aiming at the shortages of basic Flower Pollination Algorithm（FPA） with slow convergence speed, low search precision and easy to fall into local optimum, a new algorithm based on dynamic global search and Cauchy mutation DCFPA is proposed. Firstly, it uses chaotic map to enhance the randomness and uniformity of the initial distribution of pollen population. Then, global average pollen position and dynamic weight reduction factor are introduced to achieve the update of individual pollen location in the process of global pollination, which can guide algorithm to correct search direction and avoid premature convergence. Finally, Cauchy mutation is used to increase the population diversity and help the algorithm to jump out of the local optimum. The simulation experiments on six classical test functions show that compared with FPA, DCFPA algorithm has better global optimization ability, and improves the convergence speed and solution accuracy of the algorithm and also has better optimization performance than those improved algorithm in related literatures.

Key words: flower pollination algorithm, chaotic map, dynamic global search, Cauchy mutation