摘要
传统进化算法在解决4个或4个以上目标的阵列天线综合问题时,出现了选择压力不足的难题。给出了一种基于双极偏好占优的阵列天线优化设计方法,该方法借助决策者根据解决实际问题的经验给出的目标值偏好,采用TOPSIS方法,比较Pareto解之间的相对贴近度值,建立了严格的非支配关系,引导种群向高的定向辐射方向图及低的零陷值靠近。为了可视化高维空间中的解集,用高维空间对角技术法对高维空间上的解进行可视化,并将该方法与现有的3种多目标优化方法进行解集质量优劣的比较与分析。仿真结果显示,该方法在解决4个以上目标的阵列天线综合问题时具有更好的收敛性以及更多的优秀解个数。
In this work, a bipolar preferences dominance based antenna arrays optimization (AAO) method was pro- posed to enhance the selection pressure of the multi-objective evolutionary algorithms (MOEAs) on the AAO problems with more than four objectives. Considering the positive preference and negative preference of the decision-makers in real-world problems, the TOPSIS method is employed to compare solutions, construct a more rigid non-domination relationship and induce the population to move towards the position with higher directional radiation pattern and lower null value. Besides, a HSDC method is applied to analyze the experimental results. In the experimental analysis, the proposed method was compared with four state-of-the-arts multi-objective optimization algorithms. The comparison results show the higher accuracy and operating efficiency of the proposed method.
出处
《计算机科学》
CSCD
北大核心
2015年第1期268-271,296,共5页
Computer Science
基金
国家自然科学基金(61379077
61070135)
浙江省自然科学基金(LZ13F020002
LY13F030010)资助
关键词
进化算法
阵列天线综合
双极偏好占优
解集可视化
Evolutionary algorithms,Synthesis of antenna array,Bipolar preferences dominance, Visualization