Modelling Risk in Highway Infrastructure Investments: Decision-Theoretic, Bayesian, and Factor Analysis Approaches

Investments in highway infrastructure are expected to achieve a satisfactory rating for avoiding cost overruns and in fulfilling their role in enhancing the sustainability of cities and regions. Despite challenges, highway infrastructure investment assessments received insufficient research attention in modelling lifecycle cost risk and the associated sustainability effectiveness. The increasing acceptance of the sustainability rating tools for evaluating highway projects implies going beyond former attention to location and design of highway infrastructure guided mainly by functional considerations. Under present high demands from the transportation planning environment, there is emphasis on characterizing risk in the first cost as well as lifecycle costs. Likewise, there is an emphasis on inclusion of sustainability factors, including effective use of resources, in evaluating project alternatives. The use of infrastructure rating tools such as ENVISION has the potential to improve the effectiveness of investment in highway projects in terms of meeting sustainability criteria, including formal recognition of the importance of lifecycle analysis. To support the application of rating tools, research is needed in modelling risk in lifecycle cost estimates, including the identification and quantification of cost overrun factors. Besides, there is a need for a methodology for joint treatment of multi-attribute criteria that encompass cost and other factors of sustainability. To go beyond the current state of knowledge, research was carried out on: - The probability models of cost overruns. - Treating risk and uncertainty in lifecycle analyses, using decision-theoretic, utility-theoretic, and Bayesian methods for evaluation of investment alternatives. - Factor Analysis of variables that characterize the causes of cost overruns and logistics regression models based on factor analysis results. Data were obtained and analyzed on actual projects that may have experienced cost overruns. Also, a questionnaire study was implemented to obtain data from transportation jurisdictions in Canada, the USA, Middle East, and Australia. Following the study of cost overrun probability models, decision-theoretic and utility-theoretic methods were developed and illustrated in the evaluation of investment alternatives while formally treating lifecycle costs and other factors of sustainability. Finally, factor analysis and associated logistic regression models were implemented for characterizing the effect of factors that cause cost overruns.