The Research Of The Model And The Method In The Estimation Of Highway Project Cost

Nowadays, with the development of the tender in our country, traditional budget method of highway project must cost much more time. To a certain extent, a new way must be put forward to quote price exactly and to increase the possibility of win bid. It is a significant practical work that if you can predict the project cost from the characters of the project you want to estimate at the base of information of a lot of projects which have been finished.Transportation Construction Cost Management Information Systems — TCCMIS can provide the data to the model of the estimation of project cost. In this paper a model has been set up which adapt to highway project cost, making use of the coalescent of the fuzzy math and neural network, according to the characters of the highway project cost. At the end, the model has been tested by using the data of a factual project. And his major work and innovation point is following:1. This paper firstly using the model of fuzzy math and neural networks in the estimation of project cost in the field of highway project cost. This paper has build a direct contact between the characters of the highway project and the project cost, thus to predict the project cost.2. In this paper, the characters of the highway cost have been thought over to choose the main project characters in the process of the model has been built. If the characters of the highway cost aren't comprehensive, the model will not manifest the particularity and difference of the highway cost. In addition, unlike former models, time factor has been adopted in this paper.3. In this paper, a great deal of work has been done to improve the precision and practicability of the forecast of the model, which mainly include the design of the style of combining the fuzzy math and neural networks, the application of the integrating networks technique and the design of the structure and parameter of the neural networks. Further more, a lot of work hasbeen done in the facet of the number of the training sample and the choose mode.