带有成长算子的遗传算法

Genetic algorithm with growing operator

提出了伪极值点的概念，举例说明了由于遗传算法随机性强使得二进制编码遗传算法极易陷入伪极值点，致使算法收敛速度缓慢的问题．设计了一种适合于二进制编码道传算法的成长算子，该算子的引入，加强了算法的方向性，有效地防止了算法陷入伪极值点，从而大大提高了算法的收敛速度该算子不要求被寻优函数连续、可微，且该算子的引入带来的计算量的增加与性能的改善比较是可接受的，最后给出使用带有成长算子的遗传算法进行一个简单的线性系统辨识和对两个遗传算法测试函数寻优的算例，并与一般遗传算法的结果进行比较，仿真结果验证了该算法的有效性。

The concept of pseudo -extremum is introduced with an example to illustrate that binary - coded genetic algorithm (GA) often stays in the pseudo - extremum for its high randomicity so that tLhe convergence is very slow.A growing operator applied to binary - coded genetic algorithm is designed to enhance the orientation and efficiently prevent the algorithm from falling into the pseudo - extremum thereby improving the convergence speed gready. The operator does not require that the function to be optimized is continuous and differentiable. The increased computation demand due to the operator is acceptable relahve to tLhe great improvement of performance. Some examples using GA with growing operator (GGA) including a simple linear system identification and two optimization of test functions of GA are given and compared with those using general GA to verify the efficiency of the GGA.