MYRIAD is a multi-criteria decision support tool (DST) developed at THALES to handle complex decision situations by supporting risk-based assessment of scenarios (situations) and possible courses of action. It aims at capturing the preferences of a decision maker regarding several alternatives on the basis of multiple and usually conflicting criteria (conflicting in the sense that they cannot be satisfied all together). One then needs to make compromise, trade-offs among those criteria.
The MYRIAD decision support approach is split from a methodological standpoint, into two parts:
- the Modeling phase: The first step consists in constructing a multi-criteria assessment function. This construction is performed from expertise. More precisely, the decision support tool aims at reproducing decisions that an operational expert would take during operation. This is done by asking to the expert (i) what is the problem to solve, (ii) what is the relevant information needed to take a decision, (iii) how to combine the relevant input information to come up with the best decision.
- the Exploitation phase: once the decision model is captured and put in a decision support tool, the decision model is run at anytime a decision is needed.
The added-values of MYRIAD are:
- Possibility to represent subtle decision strategies of the decision maker through interacting criteria. We use very versatile decision model to represent these decision strategies;
- Advanced learning of the parameters of the decision model (weights, utilities) from intuitive learning examples;
- Automatic explanation of the assessment proposed by the decision models.
MYRIAD within PREDICT
As a one of the key components of the PREDICT Tool Suite, MYRIAD will help the end-users comparing different scenarios (situations) and assesing their consequence on the different previously mentioned operational criteria.
MYRIAD will take as inputs the results of the prediction tools (PROCeed and SBR). In the MYRIAD approach, risk is assessed using the likelihood of a scenario and its impact on different criteria and sub-criteria, such as:
- Crisis management efficiency such as the quality of evacuation, or the victims of the crisis management team;
- Vital needs such as casualties, or access of citizens to water, food or housing;
- Infrastructures such as roads, railway, aviation status;
- Livelihoods such as financial loss ; and on cross-cutting issues such as environmental issues).