Risk modelling

For large and complex system problems, it is often helpful to develop a model of the system to better understand how risks can be realised. Risk Solutions is expert in a wide range of modelling techniques. These fall into two groups:

Statistical and econometric methods to predict the impact of changes based on historical experience
Holistic modelling techniques based on a theoretical understanding of the risks involved.

The limitations of available data and sophistication of the questions being considered often means that a rigorous statistical approach is not feasible and holistic techniques are needed to model some (or all) of the problem.
Risk Solutions has wide experience of all these techniques, including:

Undertaking statistical analysis of large data sources (e.g. correlation and regression analyses, cluster analysis)
Developing Belief Networks – effectively a quantitative version of an influence diagram in which the state of the ‘nodes’ depends on that of any related nodes
Fault and event trees to model safety and reliability risk in detail
Bespoke models such as Exodis(TM) which models disease spread (such as foot and mouth disease and avian influenza) and the effect of different control strategies.

Depending on the complexity of the problem and the client’s needs, our models may be developed using Microsoft Excel (possibly using ‘add ins’ such as @Risk), specialist software or be developed in a programming language.

Whatever the approach, a key challenge is being able to cope with an imperfect understanding of the problem, which can give rise to strongly held and opposing views on how to manage the risk. Risk Solutions is adept at eliciting expert opinion on the probability of different events and the relationships that need to be modelled. In this way we are able to establish a model of the problem that has widespread support and, where there is disagreement, test the significance of this and (if necessary) model the uncertainty involved.

EXAMPLE CASE STUDIES: