Thesis Type: Postgraduate
Institution Of The Thesis: Middle East Technical University, Faculty of Engineering, Department of Civil Engineering, Turkey
Approval Date: 2019
Student: Özgür Barış
Supervisor: RİFAT SÖNMEZAbstract:
The estimation of project duration is important in construction projects, since it directly affects the costs and the success of the projects. Schedule risks can be determined along with the duration estimations in order to assess the possibility of delay penalties and additional costs. Therefore, the accuracy of the estimation of project duration range is very crucial in construction industry to assess the schedule risks. This thesis presents an integrated approach of non-parametric bootstrap sampling method and regression analysis in order to determine the project duration ranges with their probabilities for construction projects. This approach combines both the regression and probabilistic methods in order to provide the range estimation of construction project’s duration with its corresponding probability. The non-parametric bootstrap sampling method, when integrated with the regression analysis, has advantages for range estimation purposes when compared to classical simulation methods such as Monte Carlo Simulation method or probabilistic methods such as Program Evaluation and Review Technique (PERT), since it requires no assumptions regarding the probability distributions and correlations of the input data. Moreover, the proposed integrated regression-bootstrap sampling scheduling method is expected to provide more accurate results than the traditional methods due to regression analysis, which is used to take the effecting factors of the productivity into account and non-parametric bootstrap sampling method, which increases the sample size by resampling the original sample without requiring any assumptions of distributions and correlations of the activities. To expose the advantages and accuracy of the new integrated regression-bootstrap sampling scheduling method, the new method is compared with the traditional probabilistic scheduling methods through two case studies. The comparisons reveal that the proposed method presents a practical non-parametric approach that provides adequate project duration range for realistic evaluation of schedule risks of construction projects.