Research On Hybrid Particle Swarm Optimization Algorithm For Nonlineap Programming Problems

Particle swarm optimization(PSO) algorithm is a kind of swarm intelligence optimization algorithm based on the flock foraging behavior. As a kind of new algorithm in the field of evolutionary algorithms, PSO algorithm has many advantages, such as simple structure, easy to implement, fast convergence and so on. It has been widely used in function optimization, power system optimization, multi-objective optimization and other fields. However, PSO has problems of premature convergence and doesn’t have the mechanism to deal with constraint condition.Thus, it is necessary to further improve and research for the algorithm in the aspects of theories and applications.Based on the research of the PSO, some improved algorithms are proposed for solving the unconstrationed optimization problems, nonlinear 0-1 programming problems and constrained optimization problems in this thesis.The performances of these algorithms are validated through numerical experiments. The main research contents of this thesis can be summarized as follows:(1) An improved particle swarm optimization algorithm based on conjugate gradient method is proposed.The algorithm effectively combines the global search ability of particle swarm optimization algorithm and the fast local search ability of conjugate gradient method, in order to overcome the slow convergence speed and low accuracy computation in basic PSO.Numerical experiment results show that the new algorithm is an effective method for solving unimodal and multi-modal function optimization problems.(2) For solving the nonlinear 0-1 programming problems, a kind of chaotic particle swarm optimization algorithm was proposed.we transforme the nonlinear 0-1 programming problems into unconstrained the nonlinear 0-1 programming problems by using penalty function method, the chaos was introduced to initialize the population and increase the diversity, the fitness variance was used to predict whether the algorithm appear premature phenomenon or not. Numerical experiments show that the algorithm is a feasible and applicable global optimization algorithm for solving the nonlinear 0-1 programming problems.(3) For constrained optimization problems, an improved particle swarm optimization algorithm was proposed. We transform the constrained optimization into unconstrained optimization problems, and the velocity equation and inertia weight were improved so that to improve the performance of particle swarm optimization algorithm to solve the complicated and nonlinear optimization problems. The experimental results of 12 test functions and two engineering optimization problems show that the improved algorithm is a feasible and efficient global optimization algorithm. Compared with the related algorithms, it indicated the new algorithm has good stability and high calculation accuracy.Finally, we make a summary of this thesis briefly, and put forward some problems worthy to be studied in the future.