Multi-agent planning and execution with risk analysis and multi-criteria optimization: the case of rescue plans to environmental and humanitarian crisis

A disaster seriously disrupts the functioning of a society and generates human, material and economic losses.
Unfortunately, some are weakly predictable such as the terrorist. But even the predictable ones, like the cyclones
whose intensity remains difficult to assess are not always easy to forecast in terms of damage. Over the past 10
years, disasters have killed more than 700,000 people worldwide, 1.4 million injured and total economic losses of
more than 1,100 billion euros. For example, on Friday, the 23rd of August 2013, Tripoli, north of Lebanon, was
hit by terrorists. Two Sunni mosques were struck by two car bombs, leaving at least 42 dead and 500 others
injured. Likewise, on January 12, 2010, an earthquake measuring 7.3 on the Richter scale (between 0 and 12)
rocked Haiti for 35 seconds. More than 220,000 people were killed, more than 330,000 were wounded and about
1.3 million were homeless. Due to the sudden occurrence of disasters, the reaction must be quick and efficient as
much as possible in order to minimize the human loss by providing help and rescue to the injured people. The
main objective is to assist and take them as fast as possible to the nearest treatment center that can treat their cases.
The presented work concerns the definition of a new decentralized model combining proactive-reactive
approaches and Multi-agent systems (M.A.S) for solving the planning problem of a terrorist attack scenario. It is
dedicated to present a PhD research work involving both The Production Engineering Laboratory in France and
the research laboratory of the Lebanese-French University of Tripoli in Lebanon