Cost-Duration-Based Lump Sum Project Selection Framework Using Stochastic Methods for Design-Bid-Build Resurfacing Projects

Abstract

The appropriate selection of procurement tools and contracting strategies is a key factor in the successful completion of construction projects and has become an important and growing area of study for both researchers and practitioners. However, most recent and concurrent research efforts undertaken in this area are focused on the implementation of alternative project delivery methods, with little attention paid to continue improving the traditional design-bid-build (DBB) contracting approach. It is a fact that the construction industry has seen a rapid increase in the use of alternative contracting methods during the last couple of decades, but it is also a fact that DBB contracting is still the most used project delivery method in the US. Thus, any efforts towards the improvement of this contracting approach would be expected to have a significant positive impact on the construction industry. This study has been aimed to contribute to the improvement of this traditional project delivery method on a specific relevant area that has great influence on the ability of public owners to successfully complete construction project; the effective selection of payment provisions in DBB contracts. This study has been conducted for the Florida Department of Transportation (FDOT) and is specifically focused on assisting this agency in the identification of DBB resurfacing projects that would offer better value-for-money if executed with lump sum (LS) payment provisions instead of using the traditional unit price (UP) compensation approach. To achieve this research objective, the author has developed a data-driven decision-making framework designed to anticipate and compare the expected cost and schedule performance of a given DBB resurfacing project under each compensation approach. The proposed decision-making framework was developed using non-linear regression techniques, Monte Carlo Simulation, and data from 86 resurfacing projects completed by FDOT between January 2015 and March 2017: 63 UP and 23 LS projects. The proposed LS project selection framework is actually the result of integrating two sub-frameworks: one to evaluate LS candidate projects based on their expected cost performance and one to evaluate the same projects from a schedule performance perspective. These frameworks produce stochastic construction cost and duration estimates in the form nomograms. Each of the two nomograms, the cost-based and duration-based nomograms, require two inputs: number of lane miles and desired confidence level set by decision-makers. These two inputs produce four outputs per nomogram: probability of having higher construction costs/duration if UP is used instead of LS; expected project cost/duration (deterministic estimate) if LS provisions are used; the worst case scenario if LS provisions are used; and the best case scenario if LS provisions are used. The worst and best case scenarios are defined in the form of cost and time savings and losses based on the desired confidence level. Finally, the study describes a Multi-Attribute Utility Theory (MAUT) model that combines the outputs from the cost- and schedule-based nomograms into an integral LS project selection framework with the ability to make trade-offs among four cost and schedule performance objectives: 1) minimize construction costs; 2) minimize construction duration; 3) maximize cost certainty; and 4) maximize schedule certainty. The MAUT model facilitates the identification of the compensation approach with the highest overall level of satisfaction of these performance objectives, which would be the approach that offers the best value-for-money for FDOT’s resurfacing projects.