The Study Of Virtual Integrated Investment Estimation Methods On Highway Projects

The preliminary investment estimate of the engineering project is the basis for the choice of the program and the financing of funds, as well as plays a crucial role for controlling the total project cost. The simplicity and hysteresis defects of the current investment estimation methods lead to the large estimation errors and low accuracy of the investment. So it is urgent to find one scientific and effective estimation method which meets the actual conditions of the project to assure the investment estimation accuracy.This article sets up the corresponding nonlinear estimation models based on Whole Life Cost and Cost-significant theory to fit the relationship among project costs and affecting factors, in accordance with the different sizes of training samples. Firstly, this article uses the attributes reduction of rough set to mine engineering data and extract the engineering characteristics of construction projects, which can overcome the subjectivity of methods used in the past to find effective engineering characteristics. Therefore, the effectiveness of this method can be proved. According to the different size of training samples, this paper uses different estimation methods. When the finished similar engineering projects have the certain amount, this paper uses fuzzy clustering method to estimate proposed project cost. And then an example is presented to show the effectiveness and feasibility of this method. When the amount of finished similar projects is large, intelligent integrated estimation method that includes rough sets neural network prediction model, ant colony neural network model and particle swarm optimization neural network prediction model should be used. First, the rough sets neural network estimation method simplifies the network input variables, and then it predicts project cost. ACO-BP and PSO-RBF is an estimation method which uses swarm intelligence to optimize neural network to get intelligent integrated estimation method. Case Simulation results prove that this method can speed up training rate, reduce the error, and improve forecasting accuracy,which shows that the algorithm is more fit engineering actual. Therefore, the effectiveness and superiority of the algorithm can be verified. On the basis of the above methods, this article uses virtual technology to create a virtual visual model of the investment program, which can show the visual image to decision makers of the investment program.