Research And Application Of PPP Infrastructure Projects Based On Genetic Optimization Of Neural Networks

The subject of this paper comes from the Science Research Project of"The research for Financing Risk Control Mode of PPP project concerning western cities", which is one of the National Nature Science Foundation project of China [70672011].Public-Private-Partnership (PPP) is a new kind of project financing mode in public infrastructure projects. Since public sectors and private sectors have different preferences, it is difficult to allocate financing risks. How to allocate risks in different participants effectively is the key factor for the guidation of the risk and profit allocation of PPP projects and for the success of the project. Infrastructure projects have various characteristics such as large investment, long concession period and large sums of participants, thus, the risk factors are more complicated than common construction projects and it is hard to control the risk. This paper studied the genetic algorithm optimization of neural networks and applied it in PPP infrastructure projects for risk prediction. Using the genetic optimization of neural networks forecasting method in the process of project financing can guide for risk prevention.This paper has accomplished the following tasks:①This paper analyzed the application of PPP financing model in public infrastructure projects. After researching the current status of the development of genetic optimization of neural networks, it analyzed the theory of risk profile, the common financing prediction methods, and the common forecasting methods during the financing process of public infrastructure projects as well as their application fields.②Based on the artificial neural network BP (Back Propagation,BP) algorithm, a method combining artificial neural network with genetic algorithm was proposed in this paper. It studied the genetic optimization of neural network algorithm, and its application to PPP financing mode in public infrastructure projects for risk prediction.③It studied the classification of risk factors based on PPP financing of public infrastructure projects, used scores taken from experts to evaluate the risk factors, and finally utilized the fuzzy comprehensive evaluation method to calculate the various kinds of risk. After the normalization of raw risk factors data, it realized the risk prediction based on the BP algorithm and the genetic optimization of neural network algorithm respectively. The effects of experimental results of two algorithms were compared in public infrastructure projects risk forecasting. The results showed that the genetic optimization algorithm for neural network model in the PPP financing of public infrastructure projects risk forecasting is better than BP algorithm.④Through the research of a practical public infrastructure projects, this paper analyzed the requirement of PPP projects, designed and implemented the main module of the prototype system. Through the sample analysis of a public infrastructure project, it verified the feasibility and rationality of the method of introducing the genetic optimization of network algorithm into the prototype system, and compared this method to traditional risk predication method.