Font Size: a A A

Extended BP Network-Based Cost Estimation Of Urban Road

Posted on:2016-01-29Degree:MasterType:Thesis
Country:ChinaCandidate:J WangFull Text:PDF
GTID:2272330482963652Subject:Architecture and civil engineering
Abstract/Summary:PDF Full Text Request
The construction of urban road is an important part of civil engineering, which affects our daily life and China’s economics. In recent years, China plays more and more attention to urban road construction and invest a lot of money in it. While compared with the emergence of a lot of urban road projects, the management does not developed accordingly. As a result, many projects get out of control, especially for the large complex projects. They are affected by many features and hard to calculate their budgets accurately, which results in the lack of money or a waste of money, and affects the development of economic in China. To control the risk in civil engineering projects, the final cost estimation plays a key role. To forecast the final cost, several approaches and models are proposed, such as BCIS approach and Monte Carlo random estimation model. While most of these methods are based on engineering experience, subjective and are difficult to calculate the final cost accurately.To address the problems mentioned above, we proposed the extended BP network based urban road cost estimation. We are focus on the two problems. The first one is how to ascertain the features that affect the cost estimation. The second one is how to enhance the accuracy and stability of BP network. For the first problem, we intend to calculate the contribution of each feature with BP network, and decide whether we keep this feature according to its affection on cost estimation. For the other one, we integrate the extended BP model to our work, and adopt a new BP model instead of traditional model and hybrid training approach to improve the estimation accuracy.At the end of this paper, an experiment is designed to test our approach. The author collects the road project information of Huanyin block in Jinan in 2013 and 2014 and divide the information into two groups-training sample and test sample. The final result shows that our method is feasible and can be used to improve the accuracy of cost estimation. In addition, our approach plays a key role in risk controlling and cost management.
Keywords/Search Tags:BP Network, Urban Road, Neural Network, Cost Estimation, Ant Colony Algorithm
PDF Full Text Request
Related items