Off-campus UNL users: To download campus access dissertations, please use the following link to log into our proxy server with your NU ID and password. When you are done browsing please remember to return to this page and log out.
Non-UNL users: Please talk to your librarian about requesting this dissertation through interlibrary loan.
Predicting construction labor productivity with Bayesian belief networks
Construction labor productivity plays an important role in labor intensive projects. Therefore, increasing construction labor productivity is a vital task to decrease a project’s cost (time). The primary goal of this research is to investigate the feasibility of developing a comprehensive causal model that can predict construction labor productivity for various project’s situations, such as existence of “Adverse Weather,” “Changes,” “Working Overtime,” etc., while considering uncertainty. It is found that Bayesian Belief Networks (BBNs) is the best approach that can model causal relationships among different factors while considering uncertainty, simultaneously. ^ Developing a BBNs model requires to extract its structure and, for each node in the network, set up a “Conditional Probability Table.” Extensive review of other scholars’ publications, regarding factors affecting construction labor productivity, allow us to extract cause-effect diagrams for each factor. These cause-effect networks are independent sub models that by applying various structures and parameters methodologies become a separate BBN. The final step of building the comprehensive model is to combine different sub models, which after 12 iterations and combining different sub models, the primary contribution of this research to the body of knowledge, which is developing the comprehensive model, is obtained. ^ The model can do a variety of queries about the effects of a single variable, or a subset of variables, on a hypothesis variable. The findings from these queries is another contribution of this research. In this research, the hypothesis variable is the probability of “High productivity.” Various sensitivity analyses on the hypothesis variable reveals that for different network’s instantiations, the effects of similar variables are not the same. Also, it shows that the “Adverse Management Systems” can decline the probability of “High productivity,” whenever a project is in its perfect conditions, more than 70%. However, when a project is in its worst conditions, it can increase the probability of “High productivity” for less than 10%. From the main variables, “Stacking of Trades” has similar effects on the hypothesis variable with less severity. This research has wonderful applicability for project managers, cost estimators, and schedulers in their decision making process regarding costs and time of projects.^
Hazrati, Ayoub, "Predicting construction labor productivity with Bayesian belief networks" (2016). ETD collection for University of Nebraska - Lincoln. AAI10090310.