Particle Swarm Optimization Algorithm

Particle Swarm Optimization (PSO) algorithm, is one kind based on the search strategy auto-adapted stochastic optimized algorithm. Because of the simplicity and efficiency of this algorithm, afterward it's obtained the widespread attention. But the research on PSO algorithm doesn't come to perfect at the present stage; the choice of hard core's parameter of PSO algorithm still has the very great dispute. Presently, many improvements to the algorithm, while enhances algorithm performance, increases the complexity of the algorithm which does not suit to solve the problems requesting high convergence rate. Therefore, making the PSO as the object of study, founding one corrective method to be able to raise the algorithm convergence rate but not to increase the complexity of algorithm is very meaningful.Because the multi-objective optimization problems are common in the practical application, it is one of the main research areas in optimization. Therefore, it has the practical significance and the scientific research value to solve the multi-objective optimization problems. Using the conventional methods to solve this kind of problems has many limits. As the development of evolution algorithm, using multi-objective evolution algorithm to solve the multi-objective optimization problems has obtained good research results. And the most representative algorithms include NSGA2,SPEA2 and MOPSO.Now, there are some correlation research results in domestic and foreign, but there are some insufficiencies in diversity and convergence of solutions. As for my works and research results based on the existing academic achievements, it is composed of two aspects. The first one is the Modified Quantum Delta-Potential-Well-based PSO which is applied to solve the single objective optimization problems. Through the experiment carried on a series of benchmark functions, the test results indicate that the algorithm compared with original algorithm has a big improvement in the convergence rate and the stability. The second one is the Crowding Distance and Dynamic Weighted Aggregation PSO for multi-objective optimization problems. And the three performance indices of CDDWA-PSO algorithm, obtained through the simulation to three typical benchmark functions, indicate that the performance of the CDDWA-PSO algorithm is good and that the algorithm can solve multi-objective optimization problems well.