Improvement Analysis And Application Research Of Flower Pollination Optimization Algorithm

The flower pollination optimization algorithm is a new heuristic optimization algorithm designed to simulate the mechanism of plant flower pollination behavior in nature.The local search process and global search process of the algorithm simulate self-pollination and heterogeneous pollination behavior respectively,and the intensity of the local search and global search of the algorithm is balanced by switching probability.The algorithm has a simple structure,strong search capability and is easy to implement.Currently,the flower pollination algorithm has been successfully applied to various complex optimization scenarios.However,as the research progresses,researchers find that the algorithm suffers from the shortcomings of late search accuracy and easy to fall into local optimality.In this paper,some deficiencies of the current flower pollination optimization algorithm are analyzed and improved,and the improved algorithm is applied to some classical optimization problems,in order to further improve the theoretical basis of the flower pollination algorithm and expand its application range.The work of this paper mainly includes the following three aspects:(1)The Tunicate Swarm Algorithm Based Differential Variation Flower Pollination Algorithm was proposed.Aiming at the shortcomings of flower pollination algorithm such as slow convergence speed and insufficient optimization accuracy,the capsule swarm optimization algorithm is simplified and introduced into the cross-pollination stage,which provides a new search method for pollen and enhances the population diversity.At the same time,in order to improve the exploitation ability of self-flower pollination,the differential mutation strategy was proposed.And the dynamic switching probability was proposed to adjust the two search modes,which improves the overall performance of the flower pollination algorithm.(2)Singer chaotic mapping was introduced in the initialization stage of the flower pollination algorithm.In order to further improve the population diversity and algorithm robustness,the opposition learning strategy was integrated into the whole process of the algorithm,then the Singer-Map Based Algorithm of Oppositelearning Flower Pollination was proposed.Combined with the background of fuzzy clustering analysis,the improvement was applied to the fuzzy C-means algorithm,which weakened the dependence of the clustering results on the initial clustering center and enhanced the ability of the pollination algorithm to solve the fuzzy clustering problem.(3)Adaptive k-neighborhood Based Flower Pollination Algorithm was proposed.The k neighborhood ring topology was introduced in the local search stage,and the normalized fitness value was proposed to adjust the k value to avoid premature convergence of the algorithm.At the same time,the adaptive step parameter was added in the global search phase to enhance the development and exploration capability of the algorithm.Finally,the improvement was applied to the parameter tuning of traditional PID controller and fractional order PID controller,which effectively improves the low efficiency and poor control effect of the traditional method.