中文摘要：

针对传统的粒子群优化算法（PSO）和差分进化算法（DE）在解决高维复杂函数易陷于局部最优、收敛较慢、精度低等缺点,提出了基于分组的PSO与DE混合算法（PSODE）。PSODE算法把种群按维数分为两组,每组的维数为原来的一半,而种群规模不变,一组由改进的PSO操作进化,另一组由DE操作进化,然后通过信息交换机制实现协同进化。与传统的PSO算法不同,新算法按一定的概率交替使用非线性改变的惯性权重和随机取值的惯性权重,平衡了算法的全局和局部搜索能力;同时采用边界变异策略有效克服了某些粒子因陷入早熟收敛而造成搜索失败的问题,并且增加了种群多样性。通过几个标准测试函数的实验结果表明,PSODE算法的优化能力、收敛精度显著提高,同时增强了全局收敛性能,能有效地避免算法的早熟收敛问题。

英文摘要：

This paper presented a hybrid algorithm based on grouped PSO and DE（ PSODE） since the traditional one trapped in local optimum easily,showed slow convergence and low accuracy in solving high-dimensional complex functions. PSODE algorithm divided the population into two groups according to the dimension,each group reduced the dimension to half of the original. It evolved the individuals of a group by improved PSO and evolved the other individuals by DE. The two groups were co-evolved during the algorithm execution by employing an information sharing mechanism. To balance the ability of global and local searching of the PSODE algorithm,the new algorithm made use of the nonlinear changing inertia weight and random inertia weight in possibilities. Furthermore,in order to increase the diversity of the population and avoid the premature problem of the algorithm,it introduced boundary mutation strategy. The experimental results of several standard test functions illustrate that PSODE algorithm can not only improve the ability of optimization and the precision of convergence significantly,but also overcome the premature problem of the algorithm effectively.