Modification And Application Of Particle Swarm Optimization Algorithm

Particle Swarm Optimization algorithm(PSO)is one of the most important swarm intelligence techniques in computational intelligence fields.This thesis made several improvement.Then the improved algorithm applied function test and parameters determination of the Dobbins Camp(BOD-DO)water quality model.The main work is as follows:(1)First,to avoid the premature phenomenon,particle swarm optimization with different evolution(DE-PSO)was put forward in this paper.This algorithm brought the superiority of the DE to the PSO and maintained the variety of the swarm by applying the intersection and the variance to the particle.Then the global search ability of this algorithm can be maintained in the late iterations;Second,to improve the precision,particle swarm optimization with simplex method(SM-PSO)was put forward in this paper.This algorithm can improve precision by applying the simplex method;Finally,hybrid particle swarm optimization algorithm(H-PSO)was put forward in this paper.This algorithm brought the superiority of the DE and simplex method to the PSO,which can maintain the variety of the swarm and improve precision at the same time.(2)The improved algorithm and the standard PSO algorithm were tested using four standard test functions.We found that H-PSO optimization results is best,DE-PSO optimization results followed,SM-PSO is slightly better than PSO algorithm in the case of 30-dimensional and 50-dimensional.Then the convergence of H-PSO under the five kinds of inertia weight strategy is analyzed.We found that the convergence of algorithm is the best under the third kind of nonlinear strategy.(3)H-PSO algorithm is applied to parameters determination of the BOD-DO water quality model.First,we found that the ? of the objective function has certain influence on the convergence rate of PSO algorithm and the convergence rate is the highest when ?=0.5;Second,we searched a set of parameters by taking the measurement data of the former six section,then compared calculated values and measured values.We found that H-PSO can effectively determine the model parameters;Finally,the different algorithms are applied to parameters determination of the BOD-DO water quality model.The results showed that the H-PSO has the best performance.