Expert quantification of uncertainties in a risk analysis for an infrastructure project

It is shown how calibration methods for expert judgement can be used to establish consensus about the uncertainties involved in a complex project risk analysis, where no consensus was present before. A case study is presented in which a project risk analysis was carried out for a large scale infrastructure project in the Netherlands. This involved the use of new and untried technology to reduce the frequency and scale of dredging operations over the long term by subtly changing the flow in a navigable river. The technology was to be applied at two separate sites. The project risk analysis was needed to identify the cheapest method of carrying out the work under the requirement that the whole project be completed with 95% probability by the year 2005. The decision problem was modelled using an influence diagram, which was quantified by expert judgement and made use of value of information techniques. Cooke's classical method was used to calibrate the experts. This is the first time that this method has been used to carry out a value of information study.