IWTA, 17th conference
21 November 2013The International Water Technology Association organized the 17th conference in Istanbul.
- The conference is hosting by Fatih University, Istanbul
- Total number of accepted papers:102
- Number of talks:5
- Number of countries participated :20
- From Egypt, Lebanon, Turkey, Yemen, Saudia Arabia, Libya, Germany, Italy, Russia, Japan, Morocco, France, USA, Pakstan, China, Kingdom of Bahrain, Tunisia, Iran, Colombia, Greece, Algeria, Iraq, Kuwit, Yemen, Australia.
During the conference, Essam A.Gooda was the chairman in one session, cochairman in another session and presented the following paper in the third session.
Abstract of Submitted Paper
Gene Expression and Mutiple Regression Models for the
Cost of Bridge and Culvert
Essam A. Gooda1 and Mohamed A. Nassar 2
1 Professor of Water Resources Engineering, Civil Engineering Dept. Beirut Arab University, Debiah, Lebanon. e.gooda@bau.edu.lb
2 Associate professor, Water and Water Structures Eng. Dept., Faculty of Eng., Zagazig University, Zagazig, nasserzagazig@yahoo.com
1 Professor of Water Resources Engineering, Civil Engineering Dept. Beirut Arab University, Debiah, Lebanon. e.gooda@bau.edu.lb
2 Associate professor, Water and Water Structures Eng. Dept., Faculty of Eng., Zagazig University, Zagazig, nasserzagazig@yahoo.com
Abstract:
The present paper is directed to the trouble of the cost comparison between the culvert and the bridge. In addition, a combined effort of cost analysis and modeling approach is presented. The predictions of the cost for both culvert and bridge are presented using Gene Expression Programming (GEP) and Multiple Linear Regression (MLR). Furthermore, The predictions of the cost using (GEP) and (MLR) are compared. The concept of using reference bridge and reference culvert is introduced. GEP approach is able to manage effectively with data gaps. Statistics and scatter plots indicate that the new approach produces acceptably results and can be used as an alternative to the MLR. GEP models predict the cost ratio for all datasets with a relatively higher accuracy (R2 is 3.0% more than MLR), higher correlation (1% more than MLR), and lower error RMSE (0.007% less than MLR).