Research And Application Of Harmony Particle Swarm Optimization Algorithm

Optimization problems generally exist in many fields such as scientific research and engineering technology.In view of the limitations of traditional optimization methods,in recent years,intelligent optimization algorithms represented by Harmony Search and Particle Swarm Optimization have obtained the rapid development and widespread application.The intelligent optimization algorithm is relatively easy to implement,but because it does not depend on the analysis of the objective function mathematical model,the algorithm is easily vulnerable to problems such as local optimal solutions during the search process.This thesis first improves the harmony search algorithm itself,and proposes a harmony search algorithm based on Levy flight.On this basis,the particle swarm optimization algorithm is introduced,and a new hybrid optimization algorithm is proposed.By combining the two algorithms and complementing each other's advantages,the optimization performance of the algorithm is further improved.Finally,the proposed algorithm is implemented through Flexible Job Shop Scheduling(FJSP)to verify the practical significance of the proposed algorithm.Specific contents are as follows:1.In order to solve the problem that in the standard harmony search algorithm,the parameters are fixed according to experience and easily fall into local extreme values,a harmony search algorithm based on Levy flight is proposed.In combination with Levy flight,a dynamic adaptive adjustment strategy of memory bank value probability,pitch adjustment probability and fine-tuning step size parameters is proposed.Broadened the search space and increased the diversity of the population.Effectively balancing global search capabilities with local search capabilities can better satisfy the search process.This prevents the algorithm from falling into a local optimum and improves the search efficiency.Finally,the proposed algorithm is verified by the benchmark function to be effective and feasible.2.In order to solve the contradiction between "exploration" and "development" of a single optimization algorithm,a particle swarm optimization algorithm based on multi-subgroup harmony search is proposed.In the proposed algorithm,first,the harmony search algorithm based on Levy flight is randomly initialized as multiple sub-populations run independently for local search.Secondly,the optimal individuals of multiple sub-populations constitute the initial population of the elite group at the top level,and the particle swarm optimization algorithm is used for searching.At the same time,information is exchanged between the sub-populations at the bottom,and the elites at the bottom and top to improve the search efficiency of the algorithm.Finally,the algorithm is tested on the benchmark function.Experimental results show that the optimization performance of the algorithm is significantly better than that of a single optimization algorithm.3.In order to widen the application neighborhood of the algorithm,the proposed algorithm is implemented in actual engineering flexible job shop scheduling.With the goal of minimizing the maximum completion time,the sequencing of the processes and the allocation of machines are reasonably arranged.At the same time,the bottom population and the top elite group are replaced by OSV-MAV for domain search to enhance local search capabilities.Experimental results prove that the proposed algorithm has practical significance.