基于遗传算法优化模糊神经网络的倒立摆智能控制

Intellectual Control of Inverted Pendulum Based on Genetic Algorithm to Optimize Fuzzy-Neural Network

提出了一种利用遗传算法来优化模糊神经网络的倒立摆智能控制,利用RBF神经网络与模糊推理过程具有函数等价性.设计了基于模糊系统的RBF网络结构。同时采用改进的遗传算法优化了神经网络的参数和权值。其中利用一种动态的交叉率和变异率.有效地加快了收敛的速度。最后,利用Matlab软件对倒立摆进行仿真.仿真结果表明.该控制具有较好的通用性和控制效果。

One type of intellectual control of the inverted pendulum which is using genetic algorithm to optimize fuzzy-neural network is presented in this paper. The RBF network structure based on fuzzy system is designed by functional equivalence between RBF neural network and fuzzy reasoning process. Simultaneously, the parameters and weights of neural network are optimized by utilizing the improved genetic algorithm. The speed of convergence is improved efficiently by using dynamic crossover and mutation. Finally, the inverted pendulum is simulated by Matlab. The simulation results show its effectiveness and availability.