Programmes for issues such as those for the geological disposal of radioactive wastes and storage of carbon dioxide are by nature extremely complex. A structured approach for making and documenting varied kinds of decision is required to support programme design and implementation. All aspects of decision-making are likely to be subject to multiple classes of uncertainty. At each programme stage, the decision-making process must be able to identify and justify key priorities for work – including research and development (R&D) - to reduce uncertainties.

Evidence Support Logic (ESL) has been developed to support such structured decision-making, and applied to varied complex projects, including radioactive waste management studies in the UK, USA and Japan, and projects concerned with carbon capture and storage. ESL is a systematic method for analysing and presenting the level of confidence that can be placed in particular judgements. A central theme of the ESL approach concerns the independent analysis of the evidence ‘for’ and ‘against’ a judgement, in order to assess both the confidence in (or against) it, and the remaining uncertainty. As such judgements are typically complex, the main or ‘root’ judgement or hypothesis is broken down into a decision tree consisting of supporting ‘child’ hypotheses. The flow of confidence for and against these child hypotheses leads to confidence in the parent depending upon carefully selected logical operators and parameters called ‘sufficiencies’, akin to weights. The confidence is then propagated according to Interval Probability Theory.

The key point is that by assessing evidence ‘for’ and ‘against’ separately the process is honest about what is not known or is uncertain. Indeed it is a powerful tool for understanding the importance of uncertainties and which should be prioritised for further investigation. It is therefore ideal for use as a tool for research prioritisation in complex development projects.

The first tree below (taken from Paulley et al., 2012) shows an evaluation of the confidence that a site is capable of hosting a geological repository. It shows a number of areas of uncertainty. The analysis tools in TESLA, as visualised in a Tornado Plot (see below), helps to understand which of these uncertainties are priorities for R&D as they are most likely to deliver valuable benefits. The second tree shows how, in this example, R&D in targeted areas leads to a significant increase in overall confidence (76% compared to 54%). ESL has delivered high-value guidance to a range of large expenditure projects for comparatively small up-front investment. The Tornado Plot below shows the sensitivity of the decision to key hypotheses and uncertainties indicating the sensitivity of the outcomes to the evaluations of evidence at the 'leaf' level.

Paulley A, Metcalfe R and Egan M (2012). Geological disposal programme design and prioritization in the face of uncertainty: use of structured evidence support logic techniques. Mineralogical Magazine, December 2012, vol 76(8), pp. 3497-3507.