Research On Multi-Resolution Modeling Of The Urban Emergency Evacuation Simulation Based On Agent-Based Modeling

According to the features of crowd risk in public place, the researchers take effective controls in order to eliminate risk and reduce the loss. Researchers need to study the emergency mechanism of the interaction among the crowds in public place. As considering the intricate links between the crowds, researchers consider the psychological factors and management factors with the spatial structure of public place to avoid or disperse the potential risk of crowd in public place. The paper which is based on the distributed and bottom-up method uses a multi-resolution modeling architecture to simulate the crowd evacuation. In the architecture, the high-resolution level models the individuals'microcosmic behavior. Though the interaction of the high-resolution level, it forms a complex phenomenon which reflects the middle (middle-resolution level) and microscopic (low-resolution level) of urban emergency evacuation. The paper integrates artificial intelligence, computer simulation technology and GIS (Geographic Information System) technology to research urban emergency evacuation simulation. At first, the current studies of the evacuation simulation model have some weakness, such as the simple model structure and the low granularity of simulation etc. In order to present a more realistic model simulation, some model focused on optimizing the behavior algorithm without considering the cost of calculation. Some studies consider from the perspective of management, such as developing, optimizing or evaluating evacuation strategies. These unilateral researches do not satisfy the requirements of present applications. This paper intends to establish a multi-resolution architecture of the evacuation simulation. By integrating the microcosmic level, middle level and macroscopic level of the architecture, it presents various resolution levels of the crowd evacuation.Second, most of the researches which are based on mathematical derivation of the mechanism do not have a deep and comprehensive understanding at the complex relationship of the system. The Agent-Based Modeling (ABM) which is proposed by social sciences research is a new intelligent method. By modeling autonomous behavior and the interaction behavior between individual, ABM generated the bottom-up qualitative change. It shows the development and evolution of the state change process, which meets the requirements of multi-resolution modeling architecture.Third, in this paper, the micro model is different from the typical micro-evacuation models, such as the cellular automata model and the social force model. The behavioral rule of the micro model establishes the complex links between the individual and the environment. Different from focusing on the emotional factors of evacuees, the micro model emphasizes on the intelligence and learning ability of the individuals'.At Last, most of evacuation simulation software applies for the building design and performance analysis. In this paper, the evacuation environment is changed from indoor to outdoor. Outdoor space environment is more complex. The contents of this paper will integrate ABM with GIS to simulation the micro, middle and macro crowd evacuation.There is still lack of integrated theoretical architecture and a mature product. The paper has the double significance of theory and practical application to research the crowd evacuation phenomenon.