Abstract
This paper reports on an automated web-based selection tool developed to assist public sector owners in selecting projects that have the greatest potential for success using design-build. At the core of this tool are five linear regression models constructed from 108 retrospective case studies. These five models predict project success in terms of budget variance, schedule variance, conformance to expectations, administrative burden, and overall satisfaction. Project to project variation is normalized through a linear weighting model created via pairwise comparisons. The result is one overall score for success. Additionally, sensitivity analysis of the regression models filters and links advice that is used to improve the potential for success using design-build. The resultant predictive tool, the Design-Build Selector (DBS), was created using a World Wide Web platform for efficient dissemination, monitoring, and maintenance.
Original language | English |
---|---|
Pages | 147-156 |
Number of pages | 10 |
State | Published - 1998 |
Event | Proceedings of the 1998 International Computing Congress on Computing in Civil Engineering - Boston, MA, USA Duration: 18 Oct 1998 → 21 Oct 1998 |
Conference
Conference | Proceedings of the 1998 International Computing Congress on Computing in Civil Engineering |
---|---|
City | Boston, MA, USA |
Period | 18/10/98 → 21/10/98 |