The evacuation efficiency of building plans is of obvious importance to the public safety. The complexity of building plans, however, makes it difficult for the efficiency evaluation. This paper presents a computation...The evacuation efficiency of building plans is of obvious importance to the public safety. The complexity of building plans, however, makes it difficult for the efficiency evaluation. This paper presents a computational model AutoEscape, which can simulate the evacuation process for any given occupant distribution in buildings. Designed as an extensible multi-level structure, the model constructs the geometry level and occupant level and establishes the interactions between levels. The GIS-based environmental analysis is realized to automatically generate the geometric representation and formulate the cognition of agents. The multi-agent based technology is employed to simulate the crowd behaviors with autonomously acting individuals. A visualization component, which provides 3D free observations for the simulation process, is developed on the platform of OGRE and integrated into the system interface in form of ActiveX control. Finally, a case study has been conducted and the results have been compared with the results of an existing model to show the reliability and capacity of AutoEscape simulation.展开更多
基金the National Eleventh Five-Year Science and Technology Development Plan (No. 2006BAK01A02)
文摘The evacuation efficiency of building plans is of obvious importance to the public safety. The complexity of building plans, however, makes it difficult for the efficiency evaluation. This paper presents a computational model AutoEscape, which can simulate the evacuation process for any given occupant distribution in buildings. Designed as an extensible multi-level structure, the model constructs the geometry level and occupant level and establishes the interactions between levels. The GIS-based environmental analysis is realized to automatically generate the geometric representation and formulate the cognition of agents. The multi-agent based technology is employed to simulate the crowd behaviors with autonomously acting individuals. A visualization component, which provides 3D free observations for the simulation process, is developed on the platform of OGRE and integrated into the system interface in form of ActiveX control. Finally, a case study has been conducted and the results have been compared with the results of an existing model to show the reliability and capacity of AutoEscape simulation.
