改进标准粒子群算法的阵列天线综合

Improved Standard PSO Algorithm Used in Array Antenna Pattern Synthesis

针对标准粒子群(Particle Swarm Optimization,PSO)算法应用于非对称副瓣形状方向图综合时,收敛速度慢和容易早熟的缺陷,提出了一种改进标准粒子群算法。该方法借助于修正Taylor综合法先得到连续口径分布,然后通过对其抽样得到粒子群初始化的基本值,对该基本值添加随机值得到PSO优化的初始粒子种群,将该种群用于PSO迭代时,采用"精英"选择思想,即用较好的粒子替代部分较差的粒子,直到满足停止条件。文中给出了运用该方法综合的两个实例,验证了其可行性,并通过多次重复试验,验证了该方法的高效性。

When standard PSO used in asymmetric sidelobe pattern synthesis of linear array, the problems of convergence slow and early-maturing always encountered. To cope with these, a new method to change the initial values of PSO is provided in this paper. The method first gets continuous aperture distribution using fixed Taylor method, then discretizing the continuous by sampling. These sampling values can be used in initializing the population of PSO algorithm. During the process of iteration, tournament selection mechanism is used. Two examples to illustrate the practicability of this method are provided. And through repeated calculation, the results show that the method is high effectiveness.