| dc.description.abstract | Proper and effective decision-making by project owners during pre-construction phases is highly critical to the successful completion of construction projects. Due to the lack or uncertainty of project information during the project development phases, many essential decisions are typically made under significant uncertainty with decision-makers' assumptions. Ideally, such assumptions need to be validated at the end of construction to improve the decision-making process of future projects. However, post-construction evaluations are currently not actively used in highway projects, and feedback loops to improve early decision making rarely exit. The purpose of this study is to develop alternative approaches to overcome the limitations above and enable continuous improvements in data-driven decision-making for project owners.
Specifically, this study presents novel data-driven approaches for enhancing project time and cost performances via three crucial topics, i.e., contractor evaluation, contract time estimation, and cost estimation, by leveraging pre-existing but underutilized historical project data. Regarding the first topic, a framework was developed to determine actual production rates of controlling activities in a project and enable the objective evaluation of contractors' past production performance using daily work report (DWR) data, considering the influence of common contractor-independent factors on production rates. Concerning the second topic, alternative approaches to construction sequencing were developed by extracting sequential patterns among construction activities from DWR data for different project work types and applying the extracted patterns for the sequencing of new projects using sequential pattern mining algorithms, statistical tests, and the network theory. Last, a multi-objective optimization-driven approach was designed to find optimal major work items and discover new knowledge and insights for cost estimating in the scoping phase for budget authorization. Each proposed framework or approach was illustrated or validated by a case study using state highway agencies' historical data.
The research findings not only contribute to the body of knowledge but also provide practical approaches to highway agencies to enhance corresponding practices with data-driven systems without collecting any additional data. Furthermore, by periodically applying the proposed approaches to more extensive or newer datasets, the systems can be updated or improved for continuous improvements. | en |
