NONLINEAR DYNAMIC ARTIFICIAL NEURAL-NETWORK MODELING USING AN INFORMATION-THEORY BASED EXPERIMENTAL-DESIGN APPROACH

In practice, model predictive control is commonly based on a dynamic b lack-box model. For linear systems, the model is frequently based on e ither a process system's impulse response or step response. For nonlin ear cases, many works have used an artificial neural network (ANN). Th e quality of the data set used to construct the ANN model is a critica l issue. In this work, we present a systematic approach for designing the data set based on information theory. Information entropy is deriv ed to identify the mutual positions among data points in all feasible regions. In addition, information enthalpy is derived to obtain a syst em's dynamic nonlinearity. Hence, the placements of the new data are d esigned on the basis of a compromise between the information entropy a nd the information enthalpy-the information free energy. Also included herein are realistic examples such as pH control. Simulation results demonstrate that the proposed approach is highly promising in terms of obtaining a reliable black-box model, such as ANN, for model predicti ve control.