摘要
提出一种基于实数编码遗传算法的方向图综合方法,采用近亲交叉回避机制和自适应交叉变异概率等,避免了基本遗传算法过早收敛的缺点。与传统的对模值和相位均进行约束的综合方法相比,本方法不仅能够有效地获取期望方向图,而且主瓣控制机制与参考点的选取无关,综合得到的方向图与期望的方向图之间的误差较小。计算机仿真表明该算法的有效性与正确性。
A novel pattern synthesis method based on the improved real coded genetic algorithm is proposed. The closed crossing avoidance and self-adapting crossover and mutation rate are used to avoid the prema-ture and slow convergent speed of the standard genetic algorithm. Compared with the traditional pattern synthe-sis algorithms which synthesize the array patterns with both desired magnitude and phase response, the presented algorithm, which synthesizes the array patterns with only desired magnitude response, is not only efficient but also independent of the reference points. Computer simulations demonstrate the effectiveness and validity of the proposed algorithm.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2008年第6期1022-1027,共6页
Systems Engineering and Electronics
基金
国家自然科学基金资助课题(60502045)
关键词
方向图综合
遗传算法(GA)
凸规划
模值
pattern synthesis
genetic algorithm
convex programming
magnitude
