Modeling dynamic systems using universal learning network

It has already been reported that the learning algorithm of a universal learning network (ULN) by forward and backward propagation is useful for modeling, managing, and controlling of large scale complicated systems such as industrial plants, economics, social and life phenomena. ULN is a network which can model and control naturally the large scale complicated systems and consists of nonlinearly operated nodes and multi-branches that may have arbitrary time delays including zero or minus ones. Therefore, ULN can be applied to many kinds of systems which are difficult to be expressed as an ordinary first order difference equation with one sampling time delay. In this paper, a new method is presented in order to optimally model a dynamic system using ULN. For the compactness of the modeling, a special filtering structure on all of the branches that cuts unnecessary branches are introduced. From simulation results, it has been clarified that by selecting an appropriate balance parameter variable one can develop a compromised model from the modeling error and compactness of the model point of view