A hybrid intelligent system for process modeling and control using a neural network and a genetic algorithm

A hybrid expert system has been developed for system modeling and optimization by using a backpropagation neural network to represent a complex system and a genetic algorithm to find the optimal control solutions. The values are self-generated and can used to reset the system process in order to maintain a prescribed desired output.; Two applications, continuous casting process and the drug dosage in organ transplant, have been applied by using the proposed system. The conventional methods, such as mathematical model or experimental model, cannot achieve satisfactory control performance. Therefore, the neural network model is constructed for modeling a complex system and the genetic algorithm model is developed for finding the optimal control variables.; Numerical results show satisfactory control performance and reveal the hybrid expert system based on genetic algorithm and neural network has the capability to optimally control for both applications, and has the potential for more applications.