Study On Evaluation System Of Competitive Construction Bidding Based On Artificial Neural Networks

With the further reform of China's building market and in order to meet international standard after china entering WTO (World Trade Organization), the execution of engineering project has been increasingly coming to the right path. Fair completion, low-consumption and high-efficiency has become the mainstream, which makes the bid invitation and bidding more and more prevailing. With a view to ensuring project quality and at the same time shortening the construction period and economizing the capital investment into construction, it is a must to award the contract to a construction team with preferable construction techniques, relatively good management and rich experience.Generally speaking, many construction companies will be involved in the bidding of one engineering project. Therefore, it is essential to choose a contractor with strong financial ability, reliable techniques and rich experience among numerous bidding companies. Presently, the general practice for bid assessment in the construction project of china is to give a comprehensive score or give a comprehensive evaluation. These methods have their own advantages and disadvantages, especially tend to be influenced by human factor. During the bid assessment process, the human subjective factors sometimes even play a key role in making the final decision for bid assessment. With this consideration, I hereby present a different way for bid assessment: construction bid invitation and bid assessment method based on radial basis function artificial neural network (ANN).This thesis first analyzes the bid invitation mechanism and bid assessment method of China, and then based on the disadvantage of current bid assessment method and taking into consideration of relevant accomplishment of current theoretical study for bid assessment, a reasonable improvement is made on the bid assessment method from the aspect of establishing bid assessment index system, data processing for bid assessment index and the establishment of artificial neural network model, etc. This thesis presents an index system which can completely reflect the comprehensive strength of the biding enterprise. Nonlinear function is used to make the index dimensionless and enable various indexes to take on the characteristics of maximum index and therefore, uniformizes the standard of comparison and is benefit for...