改进的遗传算法及其在电磁场全局优化问题中的应用

An improved genetic algorithm and its application in global optimization of electromagnetic field problems

提出了一种适用于电磁场优化设计的改进遗传算法。分析了标准遗传算法收敛速度慢的原 因，针对其不足提出了相应的改进方案。改进算法中引入了迭代过程的分类（随机搜索阶段与局部 优化阶段）、新的群体产生规则、适应性的变异概率、新的迭代终止准则等方法。将改进算法与标准算 法用于对典型数学函数和应用实例的求解，运算结果表明，该算法比标准遗传算法不仅求解质量高， 而且收敛速度快、稳定性强。

An improved genetic algorithm for global optimization of electromagnetic field roblems isdeveloped in this paper. The reason for the slow convergence of the standard genetic algorithm is analyzed at first, then the improved genetic algorithm is proposed to expeditethe convergence. Techniques as partition of optimization process(random search stage and local optimization stage), novel criterion for new generation creation, adaptive updating of mutation probability and new stop criterion are introduced in the improved algorithm. The numeral results on a mathematical test and a practical problem both show the proposed method not only obtains better results, but also is timesaving and more robust compared with the standard genetic algorithm.