In this paper we discuss the theory of the macro-economic model and the econometricas, we also briefly present the developing history of the macro-economic model and its applications. Then we introduce the idea that the nonlinear factors make the macro- economic system operations appear to be complex and changeable. In order to set up a high performing macro-economic model, we must consider the nonlinear factors in the macro- economic systems deeply. In this paper we introduce two kinds of neural networks that are often used in many fields, one is BF' neural networks, and the other is Hopfield neural networks. We discuss the structures and characteristics of these two kinds of neural networks. After that, we construct a nonlinear macro-economic model (a brief model) with a BP neural network, then some forecasts in the macro-economics are made with this network model, and some judgments on the macro-economic policies are also made. The results of this work show that the nonlinear macro-economic model established with neural networks is better in accordance with the actual macro-economics than some linear macro-economic models that are set up with traditional methods, and this kind of macro-economic model can also be judged by some statistic judgments. The procedure of executing economic policies is the procedure of affecting and controlling the operation of economics. The establishment of the nonlinear macro-economic model is helpful very much for devising a macro-economic control system that possesses high performance, but traditional design methods of linear controller with a nonlinear macro- economic system cannot get good control results. A self-adaptive control system devised with neural networks can get good control results on the condition that the structure of controlled systems are unknown, this characteristic of the neural network controller is particularly adaptive to the macro-economic control system, so we present several kinds of the macro- economic control systems designed with neural networks, and we successfully solve the optimal programming and optimal control problems in the macro-economics with neural networks. In this paper we combine the fuzzy control theory with neural network theory, make them help each other. A brief fuzzy macro-economic model is established with this method. Then we discuss the fuzzy control system designed with neural networks. Because of the fact that the man can accumulate the system knowledge and improve the control rules gradually in the process of operating the control systems, we present a network fuzzy controller that can turn the parameters itself according to the situation of control system. The results of computer simulation show that this kind of fuzzy control system possesses potent control abilities for some systems that are unstable or unknown in structure inside. In the last part of this paper, we discuss some complicated mov
修改评论