Applying New Uncertainty Related Theories and Multicriteria Decision Analysis Methods to Snow Avalanche Risk Management

Taking the “best” decision is an important issue in the natural risks expertise and management process because of lack of information and knowledge about natural phenomena and also because of the heterogeneity and reliability of information sources (historical data, field measures, expert judgements). One main problem is therefore to help decision making with regard to the respective quality, importance, reliability of available information. This article presents a new method called ER-MCDA to help decision making considering information imperfections arising from several of more or less reliable and possibly conflicting sources of information. First, principles of existing methods are reminded. Classical methods of multi-criteria decision making and existing theories to represent and propagate information imperfections are described. In a second point, we describe the principle of a the ER-MCDA method mixing from one hand multicriteria decision analysis (MCDA) to model the decision problem and from the other hand Fuzzy Sets theory, Possibility and Evidence Theories (ER) to represent, fuse and propagate information imperfections. Experts, considered as more or less reliable, provide imprecise and uncertain evaluations of quantitative and qualitative criteria which are combined through information fusion. As an example, the method is applied to a simplified version of an existing system aiming to evaluate the sensitivity of avalanche sites. This new method allows to consider both information importance and reliability in the decision process. It also contributes to the traceability improvement. Some others developments are under progress to deal with other expertise problems such as avalanche triggering conditions or data quality.