Research On Improvement And Application Of Particle Swarm Optimization

The intelligent optimization algorithm is developed by simulating biological behavior or natural phenomena.It does not need to establish a detailed mathematical model of the problem to be optimized,thus bringing a new solution for the optimization of complex problems.As a class of intelligent algorithms,particle swarm optimization(PSO)has the advantages of simple structure,few parameters and high performance,and has been widely used in the field of optimization calculation.At the same time,the algorithm also has the disadvantages of poor population diversity and easy to fall into the local optimal solution when the program is running,which affects the convergence performance of the algorithm.In order to solve these shortcomings of the algorithm,this thesis proposes two improved strategies based on in-depth study of algorithm principle and convergence characteristics.These two strategies improve the shortcomings of the particle swarm optimization algorithm,such as low population diversity and easy to fall into the local optimal solution,and improve the global search ability and local search ability of the algorithm.An improved particle swarm optimization algorithm based on differential learning mutation(LDPSO)is proposed.In the LDPSO algorithm,the idea of DE/current-to-best/1differential mutation is combined,and a learning mutation strategy for local optimal solution of population is proposed.This strategy changes the direction of population search when the population falls into local optimal solution,so that the population can effectively jump out of the current search region and reach a new region and search again.On this basis,the algorithm incorporates the natural selection mechanism to improve the convergence performance.Through repeated experiments and comparative analysis on the test function of the algorithm,the results show that the LDPSO algorithm has better convergence accuracy than the traditional improved particle swarm optimization algorithm.A bimodal learning particle swarm optimization(BLPSO)algorithm is proposed.In BLPSO algorithm,a new position update formula is used to replace the position update formula of standard particle swarm optimization for the poorer particles in the population,but the inertia information of the individual is retained,so that the poorer particles in the population can jump to a new search area according to the position information of the center ofthe excellent particles.The information contribution of the poorer particles in the population is greatly improved and the diversity of the population is increased by re-searching the domain.By incorporating the differential mutation strategy,the fused particle swarm optimization algorithm is repeatedly tested and compared on the algorithm test function.The results show that the algorithm has better performance in the optimization of high-dimensional and complex problems,and proves the effectiveness of the strategy in the improvement of the algorithm.In a new type of controllable excitation linear motor,in order to design the corresponding controller,it is necessary to obtain the accurate parameter value of the motor during operation.The LDPSO algorithm is used to identify the parameter of the motor.By designing the appropriate ideal motor model,the fitness function is constructed.Three improved particle swarm optimization algorithms are selected for comparative analysis.The simulation results show that the LDPSO algorithm has better convergence performance in parameter identification of the motor,which is more than twice the other three algorithms,and has higher convergence accuracy with the lowest error of 0.3%.