A Design of EA-based Self-Organizing Polynomial Neural Networks using Evolutionary Algorithm for Nonlinear System Modeling

We discuss a new design methodology of self-organizing approximator technique (selforganizing polynomial neural networks (SOPNN)) using evolutionary algorithm (EA). The SOPNN dwells on the ideas of group method of data handling. The performances of SOPNN depend strongly on the number of input variables available to the model, the number of input variables and type (order) of the polynomials to each node. They must be fixed by designer in advance before the architecture is constructed. So the trial and error method must go with heavy computation burden and low efficiency. Moreover it does not guarantee that the obtained SOPNN is the best one. In this paper, we propose EAbased SOPNN to alleviate these problems. The order of the polynomial, the number of input variables, and the optimum input variables are encoded as a chromosome and fitness of each chromosome is computed. So the appropriate information of each node is evolved accordingly and tuned gradually throughout the EA iterations. We can show that the EA-based SOPNN is a sophisticated and versatile architecture which can construct models for limited data set as well as poorly defined complex problems. Comprehensive comparisons show that the performance of the EA-based SOPNN is significantly improved in the sense of approximation and prediction abilities with a much simpler structure compared with the conventional SOPNN model as well as previous identification methods.