DocumentCode
2899352
Title
The Application of Support Vector Machine on Rate of Nonlinear Direct Cost
Author
Liu, Lixia
Author_Institution
Dept. of Commun. Eng., Eng. Coll. of Armed Police Force, Xian, China
fYear
2009
fDate
7-8 Nov. 2009
Firstpage
811
Lastpage
814
Abstract
Time-cost optimization problem is one of the most important aspects of construction project planning and control. And the relationship between direct cost and activity duration is the base of project planning and control. Nowadays, when construction planners made an optimization of time-cost, they assumed the rate of direct cost is linear in order to calculate it simply. But in fact the direct cost of the project in function is non-linear. In order to obtain the relationship between direct cost and activity duration, this paper developed a universal non-linear function of the direct cost based on support vector machine (SVM). We can get the non-linear function of the direct cost through determined the coefficients of the function with SVM. This non-linear function of direct cost can describe the relationship between direct cost and duration more precise. And it can help construction planners to control project cost. After simulation of project, the validity and practicality of the method have been verified.
Keywords
construction industry; costing; optimisation; production planning; project management; support vector machines; activity duration; construction project control; construction project planning; nonlinear direct cost; nonlinear function; support vector machine; time-cost optimization problem; Communication system control; Control systems; Cost function; Educational institutions; Electronic mail; Force control; Information systems; Nonlinear control systems; Subcontracting; Support vector machines; direct cost; non-linear; support vector machine(SVM); time-cost optimization;
fLanguage
English
Publisher
ieee
Conference_Titel
Web Information Systems and Mining, 2009. WISM 2009. International Conference on
Conference_Location
Shanghai
Print_ISBN
978-0-7695-3817-4
Type
conf
DOI
10.1109/WISM.2009.168
Filename
5368409
Link To Document