EXPERIMENT DESIGN CONSIDERATIONS FOR NONLINEAR-SYSTEM IDENTIFICATION USING NEURAL NETWORKS

Although the non-linear modelling capability of neural networks is wid ely accepted there remain many issues to be addressed relating to the design of a successful identification experiment. In particular, the c hoices of process excitation signal, data sample time and neural netwo rk model structure all contribute to the success, or failure, of a neu ral network's ability to reliably approximate the dynamic behaviour of a process. This paper examines the effects of these design considerat ions in an application of a multi-layered perceptron neural network to identifying the non-linear dynamics of a simulated pH process. The im portance of identification experiment design for obtaining a network c apable of both accurate single step and long range predictions is illu strated. The use of model parsimony indices, model validation tests an d histogram analysis of training data for design of a neural network i dentification experiment are investigated. Copyright (C) 1996 Elsevier Science Ltd