Decision making and uncertainty analysis in success of construction projects

This study identifies key construction cost and schedule performance determinants related to project general characteristics, specific features, and construction best practices for heavy industrial projects. This research examines the relationship between factors corresponding to each phase of a project and develops a qualitative model to help project managers and owners ascertain project success probability at the early stages of a project. To carry out the designed research methodology, the CII-RT305 data set was used, and missing data points were generated through mean value substitution and transformed to their corresponding z-values. Several statistical tests including two sample t-test, Kruskal-Wallis test and chi-squared test were conducted to identify critical cost and schedule performance indicators. The results of the correlation analysis between project characteristics and phase-based project cost and schedule overrun are also presented. The outcomes of this analysis are used for the sequential variable reduction. The output of this screening phase is used as an input for stepwise data reduction in order to further decrease the number of potential indicators. Next, construction experience is used to incorporate the excluded cost and schedule performance indicators, if it is believed that the variable was excluded through the statistics. Then, the all-possible combination regression is used to finalize the phase-based cost and schedule performance indicators. Leamer’s and Sala-i-Martin Extreme Bounds Analysis (EBA) methods are used to study the robustness or fragility of the identified variables. In practice, the purpose of identifying robust cost and schedule performance indicators during engineering/design, procurement and construction phases is to guide project managers in allocating their limited human and machinery resources more effectively and efficiently. This research contributes to the field of construction engineering and management in two major ways. First, it identifies project factors/characteristics which drive poor project cost and schedule performance during the engineering/design, procurement and construction phases. Secondly, it determines the robustness of each of these cost and schedule performance indicators during the engineering/design, procurement and construction phases, which assists project managers to allocate their resources more effectively. For future studies, the author recommends that the coupled impact of project cost and schedule performance be studied. For this purpose, it is suggested to define the success of a project by integrating the project total cost and schedule and make a new project parameter. Categorizing the continuous data makes it possible to integrate the cost and schedule performance and develop a predictive logistic regression model to predict project success level during the conception phase.