Probabilistic modelling: theory and practice

Probabilistic modelling techniques allow much more realistic estimates of exposure and risk by computing the use of the full range of potential exposures rather than single 'worst case' exposures. However, these techniques require additional considerations regarding the appropriate data and models. This article reviews the theoretical aspects of probabilistic modelling and also considers some of the practical applications. The most common method, called Monte Carlo analysis, is discussed in some detail. The practical application of Monte Carlo to risk assessments is presented along with an evaluation of the input parameters. Topics also discussed include considerations of the requirements for precision and procedures for validation of assessments.