Risk and Reliability Analysis of Geometric Design Criteria: A Critical Synthesis

Abstract

Highway and street designers deal with the challenge of designing for a broad range of driver, vehicle, and roadway characteristics and conditions. There is significant variability in design inputs and design controls that influence design criteria and design decisions. This variability has traditionally been considered implicitly through selected values for geometric design parameters and criteria. Average values are used when the variability in the input design parameters is small. Conservative values are used if the variability is larger, often the case within the highway geometric design context. Previous research has demonstrated that addressing this variability and uncertainty more explicitly as part of design decisions can lead to better-informed and more cost-effective design decisions. Probabilistic design approaches that quantify both risk and reliability have been successfully incorporated into other design disciplines for those reasons. These approaches have also been explored in the highway geometric design literature and have shown promise. They are likely to be central to future performance-based design initiatives, as outlined in a recently published framework on conducting performance-based geometric design analysis. Given the emerging importance of performance-based design and the need to address challenges regarding the current method of handling variability and uncertainty in the input design parameters, this paper presents a collective review and assessment of methodological alternatives for quantifying risk and reliability associated with geometric design criteria and decisions.