Dynamic Modeling Method Based On Rough Sets And Rbf Network

With the rapid development of social economy, the industrial control plants are more complicated and the demand is higher. Most control systems are characterized by multi-variables, high non-linearity, tight coupling etc, and the demand of holistic control performance is high, so the demand for the complicated dynamic system modeling proposed to be much higher. Rough set theory, putted forward by Polish scientist Z. Pawlak, is a new valid mathematical tool to deal with imprecise, uncertain and incomplete data. RBF neural network is a new developed feedforward neural network in recent years. Because of its simple structure, swift training and the capabilities of approaching any nonlinear functions precisely, it has become a new tool used in system modeling. But for complicated system, it is difficult to guarantee the integrality and validity of knowledge in the initial training data when setting up the model by neural networks. There are often redundancy and noise in the training data, which make the structured network very big and the computation required may be too heavy. Thus it is difficult to reach our anticipant precision. There is fine complementarity between Rough set theory and neural networks, so integrating both of them can offer a powerful way for processing uncertain, incomplete information, which has payed more attention by more and more domestic and foreign scholars.In this dissertation, dynamic modeling method for complicated nonlinear system based on rough set theory and RBF network is researched mainly. The major innovations in this article are as follows: ? Method for Discretization of Continuous AttributesRough set theory can only deal with the discrete attributes, but it cannot deal with continuous attributes directly, this defect has limited its application range greatly. Therefore, to be discrete for continuous attributes is one of the key problems in rough set theory. As to traditional methods, the most are to process the decision table that condition attributes are continuous but decision attributes are discrete, so it is impossible for these methods applying to complicated continuous system modeling directly. In this case, this article Investigated one kindof discretization problem about decision table, which both condition and decision attributes change continuously; Discretize the condition and decision attributes respectively according to their characteristic. An algorithm for discretization based on Particle swarm optimization (PSO) is presented, which can settle the problem of continuous attributes discretization in systema modeling perfectly.? Algorithms for Core and ReductionThe rough set theory is deeply investigated, and some useful properties of the positive region are discovered. Present a method for core directly based on the positive region. And then, two algorithms for reduction based on positive region are given.? Constructing RBF Neural Networks by Rough Set TheoryUsing the advantage of rough set theory in data processing, we can get the certain rules from the training data. Every rule represents a certain class of the batch of data. If regarding the condition of the rule as input and the decision as output, each rule can be seen as a certain input-output sample. After analyzed the characteristic of the Radial basic function and RBF neural network structure in this article, using rough set theory to choose center vectors of RBF neural networks, puts forward a method for constructing RBF networks by rough set theory.? Dynamic Modeling Method Based on Rough Set Theory and RBF Neural NetworksBy virtue of respective advantage about rough set theory and RBF networks as well as their complementarity, present a dynamic modeling method based on rough set theory and RBF neural networks, introduces the process of modeling in detail. The simulating results verified the validity and superiority of this method.Zhang Tengfei (Control Theory and Control Engineering)Directed by Xiao Jianmei...