Predictive Control Based On Neural Network And Genetic Algorithm

Based on the thought of control after prediction, predictive control is obviously superior to feedback control system. With characteristics of parallel processing and strong robustness, neural network (NN) has got acclamation in control area. Genetic algorithm (GA) provides a common framework for solving optimization problems in complex system. A control method combining the three algorithms is put forward in this paper after analysing basic theory and typical algorithms. As predictive control has no form restrictions on prediction model, NN can be introduced to do prediction or optimization calculation. Taking into account that it's difficult to determine the value of neural network parameters and control parameters, GA is used to optimize parameters. Simulation application in primary steam temperature control system shows the method can not only ensure control performance, but also has strong robustness and adaptability.