Research On Improved Particle Swarm Optimization Algorith

Particle swarm optimization is a swarm intelligence algorithm that simulates biological evolution.It has the characteristics of fewer parameters and fast convergence speed.At present,when solving some optimization problems,the algorithm has problems such as low convergence accuracy and easy to fall into local optimum,which require further research and improvement on the optimization algorithm.Therefore,many scholars have proposed many improved particle swarm optimization algorithms based on the existing achievements.Combined with the above situation,several improved particle swarm optimization algorithms are proposed from different aspects of studying the theory and structure of the particle swarm optimization algorithm.The main research contents are as follows:To improve the convergence accuracy of the algorithm,a particle swarm optimization based on average position learning is proposed.The algorithm divides the entire search process into two stages,and different learning strategies are used in the preceding and following stages.In each evaluation process,the outstanding individual in the search process is selected through double competition,and it is stored in the archive.In terms of learning strategy,the average position of all particles in the population is introduced in the early stage,and the average position of all individuals in the archive is introduced in the later stage,which provides a new search direction for individual optimization.To improve the quality of non-dominated solutions in the archive,and then improve the performance of the algorithm,a hybrid multi-objective particle swarm optimization based on central control strategy is proposed.The algorithm introduces a disturbance strategy of trending value,which can improve the diversity of the population.To maintain the size of the external archive,a central control strategy is proposed to update the non-dominated solutions in the archive.Then,a strategy for updating the individual historical optimal position based on combination method is introduced to maintain population diversity.To utilize the computational cost effectively,a multi-objective particle swarm optimization for maintaining dynamic population size is proposed.The algorithm proposes an improved individual historical optimal position and population historical optimal position update strategy to alleviate the population from falling into local optimum.Then,a strategy for maintaining dynamic population size is proposed,which can enhance the quality of particles in the population and avoid excessive population size.The improved algorithms is compared with the existing optimization algorithms in solving the selected test function.The results verify the effectiveness of the improved algorithms.