Nonlinear system identification using genetic programming

Abstract

The objective of the present study is to develop system identification tools using genetic program paradigm. The developed software is able to identify several types of models ranging from algebraic to differential equation system using process data. Identification of state space model using genetic programming is a relatively new area of application that has been attempted here. A unique approach in model representation, which helps to develop faster and robust computer programs, has been attempted. We introduce a new concept called a??evolution policya?? to take advantage of process knowledge. A new fitness measure that takes into account functional complexity of the model is also proposed. To improve the efficiency of genetic programming, we also propose enhancements such as modified genetic operators, new block model representation using Simulink process simulator, distributed computing, integration of nonparametric techniques and implicit algebraic equation modeling.