论文部分内容阅读
采用多层前馈遗传神经网络模型对甘蔗制糖结晶速度进行学习和预测 ,并针对该模型存在的计算量大 ,收敛慢的问题 ,采用具有强化作用的Q学习确定遗传算法的变异概率 ,以提高学习的收敛速度 ,仿真结果表明了该方法的有效性
The multi-layer feedforward genetic neural network model is used to study and predict the sugar crystallization speed of sugar cane. In order to solve the problem of large amount of calculation and slow convergence of the model, Q mutation with genetic algorithm is used to determine the mutation probability of genetic algorithm Improve the convergence rate of learning, simulation results show the effectiveness of the method