The subject of this thesis is risk analysis and decision support in the context of transport infrastructure assessment. During my research I have observed a tendency in studies of assessing transport projects of overlooking the substantial amount of uncertainties within the decision making process. Even though vast amounts of money are spent upon preliminary models, environmental investigations, public hearings, etc., the resulting outcome is given by point estimates, i.e. in terms of net present values or benefit-cost rates. This thesis highlights the perspective of risks when assessing transport projects, namely by moving from point estimates to interval results. The main focus of this Ph.D. study has been to develop a valid, flexible and functional decision support tool in which risk oriented aspects of project evaluation is implemented. Throughout the study six papers have been produced laying the foundation with different case examples ranging from road, rail to air transport projects. Two major concerns in building the assessment model, CBA-DK, are to bring informed decision support to the decision-makers and to specify relevant probability distribution functions to feed into the Monte Carlo simulation, being the technique behind the quantitative risk analysis of CBA-DK. The informed decision support is dealt with by a set of resulting accumulated descending graphs (ADG) which makes it possible for decision-makers to come to terms with their risk aversion given a specific decision task. ADG depicts the decision-makers risk aversion towards a specific assessment task, i.e. by illustrating probabilities of an infeasible socio-economic rate of return. To perform informed decision support as proposed by ADG it is necessary to determine a set of suitable probability distributions. This selection process has been conducted among others by literature studies, conference and seminar attendances and substantial amount of tests within CBA-DK. Currently, the model is made up by five different distributions further divided into two groups of non-parametric and parametric functions. New research proved that specifically two impacts stood out in transport project assessment, namely, travel time savings and construction costs. The final concern of this study has been the fitting of distributions, e.g. by the use of data from major databases developed in which Optimism Bias and Reference Class Forecasting are implemented. Throughout the entire research from the beginning in 2004 to this day, the modelling framework of CBA-DK has evolved and changed radically. Recently, Palisade Corporation, the developer of @RISK, issued the new version 5.0 allowing for a much greater freedom when choosing probability distributions and performing real term data fits. The perspective of this Ph.D. study presents newer and better understanding of assigning risks within assessment of transport projects.
Monte Carlo simulation; Cost-benefit analysis; Risk analysis; Transport assessment; Decision Support Model