中文摘要：

针对粒子群算法在寻优过程中存在的容易陷入局部极小、收敛速度慢等缺点,结合粒子在实际寻找食物的过程中,大部分可以飞到其预期的最佳位置,而少数粒子由于受不确定因素影响,发生飞行偏离,本文提出了一种改进粒子群算法。算法中的模拟不确定因素干扰操作,能够有效避免群体过度集中现象,有效增加了种群的多样性。典型复杂机械优化设计的仿真结果表明,该改进算法能够快速、有效地进行全局搜索。

英文摘要：

To overcome the shortcomings of particle swarm algorithm （PSA） such as premature convergence and/or slow speed of convergence, we propose an improved PSA to guarantee the convergence and obtain global optimization solutions for complex problems. Most particles in the PSA can fly to their anticipated best positions in the process of searching for food, but some particles may depart from ideal routines owing to uncertain factors. The algorithm can avoid the over-concentration of particle swarms and effectively increase the diversity of their species. Simulation results of the optimal design of typical and complex machines show that the improved algorithm can effectively perform global search.