The Research Of A Computational Model For Large-Scale Open Space Emergency Evacuation

Decision-maker used to depend on the knowledge, interest, cognition, experience,creativity and sense, because the complexity, limited resources and time in thelarge-scale open space emergency evacuation. And completely rationaldecision-making does not apply. Emergency evacuation decision-making can beinvolved in the physical domain, the information domain and cognitive domains basedon the extended OODA loop decision-making model. Therefore, the prediction can beobtained though perceive the situation. Corresponding, the decision-maker candevelop the emergency evacuation plan and execute it. This paper focuses on acomputational model for large-scale open space emergency evacuation based on theextended OODA loop model. The details of results are as follows:(1) The BDIP Agent model which increases the prediction function based on theBDI Agent model was designed. Accordingly, this paper presents a formalisation ofpredictable element within the BDI model proposed by Rao and Georgeff andindicates the extended axiom, semantics and decision-making behavior. And thesoundness and the completeness of the BDIP model are proved. Corresponding theprocess and algorithm of task confirmation is also given. Consequently, the problemof task confirmation can be sloved base on the BDIP Agent model.(2) The emergency evacuation task allocation model called EV-GAP wasproposed. First of all, the formal definition of the emergency evacuation taskallocation problem is given. And the emergency evacuation Agent’s ability wasdefined. The prediction about system feedback can be obtained by using the futuredata of the population and the speed. Moreover, the task allocation algorithm andgeneration method of emergency evacuation was given. Finally, the effectiveness ofthe model has been verified through emergency evacuation simulation.(3) The emergency collaboration Petri net model focuses on the basic structure ofAgent and Multi-Agent collaboration is proposed. Correspondingly, the modelingmethod and collaboration prediction algorithm of emergency collaboration Petri netmodel is also given. In addition to describe the respond of the emergency evacuationprocess, this model can be used to predict the conflict between tasks in the emergencyevacuation plans to solve the problem of collaboration during emergency evacuationprocess. Finally, an emergency evacuation decision-making prototype system based onabove research was introduced to verify our results.