摘要： 针对传统粒子群优化算法容易早熟、收敛精度低等缺点，提出一种改进方案，使用随机惯性权重，在每一次迭代中，对可能陷入局部极值的粒子进行有效的随机初始化。通过对7个经典测试函数的数值仿真实验证明，该新算法能提高粒子群优化算法的寻优能力，并在维数较高时也能获得较好的优化效果。

关键词: 粒子群优化, 权重, 随机初始化

Abstract: To overcome the disadvantages of Particle Swarm Optimization(PSO) algorithm such as premature, bad convergence precision, a new improvement scheme is presented. In the new algorithm, random inertia weight is used, and at each iteration, the particles which may fall into the local extremum area are randomly initialized renewedly in an effective method. The numeric experiments of seven classic benchmark functions indicate that the new algorithm greatly improves the particles’ ability of searching global optimal value, and has good results in higher dimensions.

Key words: Particle Swarm Optimization(PSO), weight, random initialization

中图分类号: 

• TP311.52