A Framework for Managing Emergency Situations via Modular BDI Architecture

In this paper, we propose a new framework based on multi-agent systems and modular BDI architecture for human and social behavior during emergency evacuation. Information flow in emergency cases is as many as in usual situations, generally irregular and lacks in reliability. It is very difficult for decision makers who are under stress to take right decisions under these circumstances. Therefore, we should consider the most important information in hand so that agents can decide better. We use social attributes addition to personal characteristics for agents in the simulation environment based on the proposed framework. The proposed framework is also able to extend similar simulations that implement BDI architecture. The separation of beliefs, desires and intentions from each other in our framework has the advantage of being modifiable. It is developed on the first framework rendering it easy to use and simple for creating bigger simulations. In comparison between the proposed method and others, our method is simulated in less time than others.Therefore, agents in our system act faster and can quickly escape from the emergency enviroment.

Multi-agent system, BDI artichecture, Modeling, Emergancy