Risk Prediction in a Space-Time Markov Model applied to Propagating Contamination

Hélčne Soubaras

Our concern is crisis situations involving propagating threats such as toxic clouds, pandemics, contamination, flows or oil slicks. Emergency planning may be then complex, and decision-making support (DS) tools based on risk criteria can be helpful. The evaluation of quantitative risk needs for a mathematical model taking into account either the cost to pay if a given point is contaminated (damage, casualties...), either the associated probability. The present work proposes original mathematical expressions for a recursive implementation of the mid-term predicted risk at each point of the threatened area. An application of the proposed risk prediction model is shown on a dangerous cloud, with numerical results.

PDF full paper