栅格法构建面向群体行为仿真的分层离散信息空间

群体行为模型运行过程中的大量搜索计算,极大制约了仿真群体规模的扩大,影响了仿真的实时可视化表现。为此提出了采用栅格法构建分层离散信息空间的解决方案,重点介绍了静态信息层、动态信息层和行为约束层的设计,并简要论述了离散信息空间中虚拟人对象运动行为的仿真实现。实验结果表明,此法能够简化群体行为模型的决策机制,有效支持群体行为模型的实时运行。

Establishment of Discrete Information-space Using Grid Method for Crowd Behavior Simulation

Lots of searching calculations of crowd behavior simulation restrict the extension of virtual crowd's scale, and also influence the model's real-time visual impression, so establishment of discrete information-space was proposed. Three layers, namely layer of static information, layer of dynamic information and layer of behavior restriction were discussed emphatically, some issues like granularity of grid, update of dynamic information and support of behavior decision-making were investigated, and realization of virtual human's motion and behavior were also brought forward briefly. Results of experiment prove that this method could predigest the decision-making mechanism and effectively support the real-time simulation of crowd behavior model.