关键词: 全局优化, 改进的混沌粒子群遗传算法, 混沌序列, 计算精度, 收敛速度

Abstract: The advantages of genetic algorithm(GA),the particle swarm optimization(PSO) and chaotic motion characteristics are combined in this paper. The chaotic particle swarm genetic algorithm (DCPSO-GA) joined with the chaos perturbing is put forward,and the global optimization performance of the hybrid algorithm are analyzed by 5 high dimensional nonlinear test function. The stagnation phenomenon which appears in the optimal search is solved by DCPSO-GA. The search space of the global optimization is expanded and the diversity of the particle is enriched,while the function gradient information is not required. The global optimal solution can be found by DCPSO-GA for the 5 test function in this paper,and its convergence rate is very fast,greatly reducing the amount of computation. Moreover,it can be known that when the total number of target function calls is close to or less than other related algorithms,the improved algorithm has a great improvement in the calculation accuracy and convergence speed. The DCPSO-GA algorithm is applied to heavy oil cracking parameter estimation and prediction. It can be shown in the test results that the parameter estimation and prediction accuracy can be improved,the error can be reduced,the global optimal solution can be effectively found,the convergence speed can be improved and the amount of calculation can be greatly reduced.

Key words: global optimization, the improved chaotic particle swarm genetic algorithm, chaotic sequence, computational precision, rate of convergence