摘要
为了克服基本遗传算法参数较多时编码表示冗长、烦杂以及实数编码等寻优效果的不足,提出了一种改进编码的遗传算法——矩阵编码遗传算法。它是在遗传算法大的框架不变的情况下仅改变其编码,即遗传编码改变的同时,相应的遗传操作,包括交叉、变异等都要随之变化。除此之外,适应度函数和计算也要变化。最小二乘法是系统辨识常用方法之一,将改进的遗传算法与最小二乘方法相结合来解决系统辨识的问题,给辨识问题的解决提供了新方法和新思路,同时也丰富了遗传算法的实际应用意义。MATLAB仿真实验结果表明,该算法可以解决系统辨识问题。
In order to overcome the def iciency and limitation of basic genetic algorithm（BGA）,the authors presentl a new GA--matrix coding GA.The method only changes the coding,needn＇t change the frame of BGA,namely choice BGA coding and operation corresponding heredity.It includes crossing,mutation along with movement.Except that adoption and calculation of fitness function and basic BGA are the same.least squares is a common ways in model identification,the improved GA combined with least-squares method to solve system identification problems,It provided a new approach and new ideas to identify problems,but also enriched the meaning of the practical application of GA.MATLAB simulation results show that the algorithm can solve the system identification problem.
出处
《机械管理开发》
2011年第5期207-208,212,共3页
Mechanical Management and Development
