遗传算法与神经控制的融合算法研究

遗传算法和神经控制是现代智能控制常用的两种方法,它们具有各自的优点和不足。将遗传算法用于前向神经网络的可能性进行了研究,阐明了遗传算法和神经网络结合的必要性。针对遗传算法和神经控制各自的优缺点,设计了二者的融合算法,将遗传算法应用于神经网络控制器的学习和训练,从而使建立的控制器兼有二者的优点,具有神经网络的广泛映射能力和遗传算法快速全局收敛以及增强式学习等性能,继而提高控制系统的性能。运用该方法对电加热炉温度控制系统进行的仿真实验,实验结果体现了良好的控制效果,证明了融合算法的优越性。

Research on syncretic algorithm of genetic algorithm and neural control

Genetic algorithm and neural control are the two common methods used in the modern intelligent control,both of which have their respective advantages and disadvantages.The paper studies the possibility of genetic algorithm to be applied in the forward neural net and illustrates the essentiality of the combination of genetic algorithm and neural net.According to the advantages and disadvantages of both genetic algorithm and neural control,a syncretic algorithm of them has been designed in this paper,in which genetic algorithm is applied to the learning and training of neural net controller in order that set-up controller is featured with the advantages of the above-mentioned two factors,that is,it has both the capacity of widespread mappings of the neural net and the performance of the genetic algorithm,such as high-speed global convergence and enhanced study etc.Therefore,it can improve the performances of the control system.This method has been used in the simulation test carried out in the temperature control system of the electric furnace and a good control effect has been achieved based on the result of the simulation,which has proved the syncretic algorithm mentioned in this article to be very superior.