| Intelligent optimization algorithms have been a popular area of research for scholars at home and abroad.They are widely used in various engineering problems due to their advantages of robustness,ease of understanding,and high solution efficiency.Many excellent intelligent optimization algorithms have been proposed,such as particle swarm algorithm,differential evolution algorithm,artificial bee colony algorithm,etc.Flower Pollination Algorithm(FPA)is a simple and efficient meta-heuristic swarm intelligence optimization algorithm proposed by Yang in 2012.The algorithm analogizes the processes of heterogeneous and self-pollination of plants to the global search and local search in the algorithm,and balances the switch between global search and local search.Due to the simple structure,few parameters,robustness and adaptability of the FPA algorithm,it has been applied to solve complex optimization problems solved in many fields.Starting from the study of flower pollination optimization algorithm,this paper further applies the improved flower pollination algorithm in different applications.The main research work and contributions are as follows:(1)A pollination algorithm based on enhanced lens imaging and neighborhood mutation strategy was proposed.Based on the reverse solution generated by standard lens imaging,the improved algorithm uses the Cartesian product between the optimal solution generated by the algorithm and the reverse solution of lens imaging to generate the elite set.The ordered pairs within the elite set are respectively calculated for the adaptation values.There are always 10 individuals in the elite set,and the new high-quality solution will replace the inferior solution.The optimization accuracy of the algorithm is further improved by expanding the set of candidate solutions.The neighborhood variation strategy takes the optimal solution of each generation as the center,and generates a new optimal solution to the information of all the solutions in the neighborhood to improve the local search ability of the algorithm.(2)A pollination algorithm based on lens imaging and traction variation was proposed.The traction mutation strategy is used to enhance the ability of the algorithm to jump out of the local optimal in the local search stage.At the beginning of the FPA mutation strategy,the exploration mode is more global search and clustering is based on the global optimum.With the increase of iteration,the algorithm takes the global optimal individual as the center for local search to seek the optimal solution.However,excessive reliance on the current optimal solution has certain blindness,and the inferior optimal solution leads to the deviation of the search direction.Therefore,traction factor is added to constrain the individual search direction,and correction is given when the algorithm search direction deviates.Increase the population diversity of lens imaging strategy optimization algorithms.An adaptive adjustment factor is added to the standard lens imaging strategy,which can control the distance between the current solution and its reverse solution.In addition,the observation factor is added to judge whether the algorithm falls into the local optimal solution.If the local optimal solution is caught,the mutation mode is changed,and the reverse solution is generated through lens imaging and the current optimal solution is updated.Otherwise,the mutation strategy is not changed and the reverse solution is not generated.For the improved FPA algorithm mentioned above,it has been applied to the optimization of automotive transmission parameters and 3D path planning problems,and experimental comparisons have shown significant improvements in both relevant parameter optimization and search speed after the algorithm has been improved. |
