Development of Fuzzy Hybrid Approaches to Project Delivery Method Selection in Highway Construction

Abstract

Selection of project delivery methods is a success factor in delivering highway construction projects because it has a substantial impact on the project performance, such as cost, time, and quality. Project delivery decision-making processes have been heavily relied on experts’ opinions and subjective judgements of professionals to evaluate quantitative and qualitative decision variables. Although current quantitative and probabilistic methods provide a robust means to analyze quantitative variables, they are not ideally suited for treating uncertainties encountered in qualitative variables. Fuzzy set theory is a mathematical approach that can accommodate a combination of quantitative and qualitative variables. This dissertation aimed at investigating the applications of fuzzy set theory and fuzzy logic to support decision-making processes in project delivery method selections. Using an empirical dataset of 254 completed highway construction projects, three fuzzy-based applications, including fuzzy cluster analysis, fuzzy pattern recognition, and fuzzy Bayesian inference system were developed, trained, and tested. As a result, fuzzy cluster analysis was used to establish seven common project clusters that share high similarities in project characteristics, project complexity, delivery risks, cost growth, and project delivery methods. Fuzzy pattern recognition was used to develop a fuzzy rule-based inference system based on the seven identified project clusters to help recognize an appropriate project delivery method associated with potential cost growth for new highway projects. Fuzzy Bayesian networks were used to develop the theoretical framework of fuzzy Bayesian inference system which is able to depict the causal relationships between project characteristics, project complexity, delivery risks, and project delivery methods. The flexibility of fuzzy membership functions in the developed applications helps leverage the evaluation of a combination of quantitative and qualitative variables in highway project delivery method selection. In addition, these data-driven fuzzy applications also allow for multiple decision scenarios based on the decision maker’s judgements of delivery risks and project complexity. This dissertation contributes to the body of knowledge by demonstrating quantitative approaches derived from fuzzy set theory and fuzzy logic to support the selection of project delivery methods in highway construction. Additionally, the results from the developed fuzzy-based applications also provide insights regarding cost performance comparisons between project delivery methods. This study may assist highway agencies in making project delivery decisions based on project attributes, historical data and their relevant experience.