【摘要】 针对最大似然(ML)方位估计全局寻优问题提出一种将文化算法用于方位估计的算法,并给出该算法的2个实现版本,即CA-version1利用规范知识调整变量变化步长,形势知识调整其变化方向;CA-version2利用规范知识调整变量变化步长及变化方向.文化算法所具有的双层结构特性,使其在问题求解过程中能够利用经验知识来指导搜索过程,从而具有较好的全局寻优性能.仿真结果表明该方法在获得与AP算法相当性能的情况下有更低的计算量.同时作为一个多进化过程的算法其结构易于并行实现,可以进一步满足实时性的要求.更多还原

【Abstract】 This paper presents a method to perform maximum likelihood estimation(MLE) for direction of arrivals(DOA).Two different versions of the implementation of the method are proposed,CA-version1 and CA-version2.The former uses situational knowledge to control changes in direction and normative knowledge to adjust the change of step size;the latter employs normative knowledge to control the changes in both step size and direction.Cultural algorithms are a dual inheritance system,so they guide the searching process by using empirical knowledge to find the global optimal solution.The simulation results showed that this method can get almost the same performance as the AP algorithm,but with much lower computational efforts.Furthermore,as an algorithm with multiple evolutionary processes,it is easy to achieve parallel operation,better meeting the requirements of real-time operation.更多还原