Workpackages

Standard for decision making in crises

Objectives

To establish a loss function loss(N) and p_N|i(N|decision_i) for integrated standard. The loss function must reflect in a balanced manner the societal concerns pertinent to a "large" loss. The (N|decision_i) will reflect the requirements on the methods to be used for generating basic information on stability, evacuation and rescue process as well as the associated uncertainty.

Description of work

Task 6.1 Develop loss function
(Responsible: SSRC, Participants: NTUA, MCA)

An exercise of prototype modelling will be undertaken to put forward various alternatives for representing the nature of the loss in terms of societal perception of the consequence and expectation of the handling of crises situations. Tentative concepts used in various risk industries, such as that the loss is a number of fatalities for a given casualty scenario loss(N)=N, or some other function, such as e.g. loss(N) = a ∙ N + b ∙ N^p, will be considered. Importance of consideration of effects such as discomfort, or possibly long-term distress mental impact or other harm will be established. An expert as well as public consultation on the proposals will be undertaken by presenting with illustrative case studies. In depth scrutiny of real life cases from the past will be undertaken, whereby the casualty mitigation actions, the eventual outcome as well as post accident conclusions will be used to commit and argue for a specific loss function. The main objective is for the loss function to represent and adequately reflect different levels of tolerability for small or big losses.

Task 6.2 Develop likelihood function
(Responsible: SSRC, Participants: NTUA, MCA)

The likelihood function is that of probability mass distribution for occurrence of specific number of fatalities or equivalent fatalities N, conditional on any particular decision being taken in specific casualty case and at given instant of time, written as p_N|i(N|decision_i). The function must combine the outcome from predictions of the process of ship stability deterioration as well as mustering, abandonment and rescue of the specific number of persons considered. The function must also represent the uncertainty in the modelling of these processes. It is expected that the function itself will be a mathematical construct rather than a numerical model, and that it will be well suited for straightforward implementation in a typical decision support systems and that it will be well suited for rapid execution if necessary. The effort will be put in Task 7.2 into setting up of appropriate schemes for relating of the information about the actual casualty case with information needed for performing the prediction.

Deliverables

• D6.1 Report on the details and the rationale of the loss function
• D6.2 Report on the details of the likelihood function