精英类克隆选择算法平均收敛速度估计

Convergence Rate Estimation of Elitist Clonal Selection Algorithm

克隆选择算法收敛速度估计是算法研究的一个难问题,目前还是处于初始的研究阶段.本文对一大类精英保持策略克隆选择算法的收敛速度问题进行了研究.首先利用算法种群中最佳个体的定向转移概率导出最佳个体的转移概率矩阵,针对实际应用中由于算法种群规模过大而导致该矩阵求取较困难的问题,将最佳个体的转移概率矩阵构造成满足一定条件的矩阵范数,从而提出一种更为简单有效的算法平均收敛速度估计的新方法.对不同的精英保持策略克隆选择算法进行了收敛速度估计仿真实验,其结果表明了该估计方法的有效性.

Convergence rate estimation of clonal selection algorithm is a difficult problem and it is still in the initial stage. The convergence rate of elitist clonal selection algorithm is studied in this paper. The best individual transition probability matrix is derived from the best individual directional transition probability in algorithm populations. It is difficult to calculate the matrix due to the large algorithm population size in practical applications. On the basis of certain conditions, the best individual transition probability matrix is conslmcted to a matrix norm and a simpler and more effective new average convergence rate estimation method of a class of clonal selection algorithm is proposed. The simulation experiments of different elitist clonal selection algorithms show the validity of the estimation method.