一种新的基于进化规划方法的盲CMA快速收敛算法

A new quickly converged CMA algorithm based on Evolutionary Programming

本文把改进的进化规划方法引进到解决均衡问题中,定义以代价函数为自变量的一种新的函数作为高斯随机化算子的标准差,来控制算法的收敛,提出了一种新的基于进化规划的CMA算法(QEPCMA),该方法应用进化规划来估计均衡器的系数,进一步提高了文献[5]中改进的CMA算法的收敛速度。通过Monte—Carlo仿真,比较了这两种算法和标准 CMA算法的性能,验证了所提算法的有效性。

Evolutionary programming（EP） is introduced into equalizations in this paper, a new function is defined which uses the cost function as independent variable. This new function is selected to be the standard variance and it can control the convergence speed of the equalization algorithm. A new algorithm named quick CMA based EP, in short, QEPCMA is proposed, where EP is used to estimate the equalizer＇ s coefficients, the algorithm＇ s convergence speed is increased comparing to the CMA and the edited CMA proposed in article[ 5 ].