Risk-Based Project Delivery Selection Model for Highway Design and Construction

Abstract

Project-delivery methods allocate risk for design and construction between contractual parties. State departments of transportation (DOTs) using federal funds employ three primary project-delivery methods: (1) design-bid-build (D-B-B); (2) design-build (D-B); and (3) construction manager/general contractor (CM/GC). Because the choice of a project-delivery method is best made early in the project-development process, it is a complex decision that is fraught with risk and uncertainty. This paper presents a risk-based modeling methodology to evaluate and quantify the potential differences in project cost attributable to the selection of a project-delivery method. The risk-based model consists of (1) an input structure of assessment and evaluation of delivery-risk factors; (2) a computational-modeling structure for calculating costs; and (3) an output structure to communicate model results and implementation. The assessment and evaluation process determines the risks that are incorporated into the delivery decision. It translates static cost and schedule uncertainty from project specifics to input variables (risk factors) and to decision variables (project outcomes). The computational model employs crossimpact analysis techniques and probabilistic inferences to capture uncertainties and interactions among the input and decision variables. The model result provides three approximate cost distributions associated with three project-delivery methods (D-B, D-B-B, and CM/GC) and a sensitivity result (i.e., tornado diagrams) that describes which risk factors have the most significant impact on these costs. The model was successfully tested on three highway projects which are discussed in detail in this paper. The findings from this paper add to the existing body of knowledge by providing a novel method to predict project costs based upon the owner’s choice of alternative project-delivery methods. The approach combines multivariate analysis with crossimpact analysis to make the predictions and provide a sensitivity analysis for project risks. The findings also provide a systematic approach to quantitatively selecting an appropriate delivery method that encourages highway agencies to conduct risk analysis early in the project-development process.