Modeling a disaster management with Decentralized Markov Decision Processes

In this paper, we propose to model the disaster mitigation, search and rescue problem with a Decentralized Markov Decision Process (DEC-MDP). This type of problem is confronted to several impediments like the time pressure, the lack of resources, ... However, the basic DEC-MDP model does not include any constraint on the execution of tasks. Thus, the aim of this work is to adapt the DEC-MDP in order to take into consideration temporal constraints, precedence constraints, uncertain durations of action execution and constraints on communication between agents.