Research Of Engineering Claims Based On The Artificial Neural Networks Model

Nowadays construction industry is facing all kinds of challenges, such as the huge building scale, the increasingly risk, the harsh construction environment and the increasingly amount of investment. As a result, that leads to the types of contracts are gradually various,and the variation of contracts is becoming more and more frequent at the same time, which are likely to cause a series of claims after the contracts are signed. Engineering claim is the inevitable objective phenomenon during the whole process of the construction in each country,therefore, it is urgent issues to tackle for each contractor how to forecast, discovery, and deal with claims effectively. Although international engineering claim is paid close attention by the project managers with a growing number of international engineering in recent years, the level of engineering claims management is still low and the knowledge about the prediction theory and practice of engineering claims is quit deficient, because our country is relatively late to research about it. It is still a difficult problem that bothers the contractor to deal with engineering claims how to predict and discover potential claim opportunity, and how to predict the result of engineering claims according to the events of claims. In view of this, it has a strong practical significance and application value to study how to effectively improve the predicting ability about the result of engineering claims, which can reduce the contracting risk and maintain their own legitimate rights and interests.By combining engineering claim theory with artificial neural networks theory, the corresponding model is constructed to effectively predict the schedule and cast of engineering claims. In terms of engineering claim theory, first of all, it is necessary to deeply study the nature of factors that lead to engineering claims form the roots of causing engineering claim,the analysis and identification of engineering claim opportunity. Then how to recognize the claim opportunity in the construction process by applying the theory of time-varying system of the contract status is studied further, the engineering claim work procedure and claim negotiation skills are introduced simultaneously. In terms of artificial neural networks theory,first of all, the artificial neural networks theory is introduced in detail in this paper, the natureof its content and the value of application are analyzed, which lay a theoretical foundation for establishing engineering claim prediction model. Then, the principle of the standard BP algorithm of artificial neural networks model is deeply analyzed about the advantages and disadvantages of its. The problem of slow convergence speed of the BP network is solved by applying the training methods LM algorithm embedded in the application of MATLAB.Finally, how to use MATLAB software finish artificial neural network prediction process is introduced.Based on the above theoretical research, eventually the LM-BP neural networks prediction model of engineering claim is built according to the organic combination of BP model and LM algorithm. The accuracy of the prediction model is verified through the engineering examples, compared with standard algorithm of BP network prediction model,finally the prediction model is applied into practice for the management of the engineering claims. Using artificial neural networks prediction, the contractor can know what factors change will cause the claims for extension of time and cast. Therefore, the contractor can collect claim evidence in advance according to the result of prediction before making engineering claim; the contractor can take timely measures to make the claim according to the process of claim work after the claim happens; the contractor can use the claim negotiation skills to minimize the loss, to gain the successful project claim.