Abstract This paper presents a mass evacuation simulator capable to handle complex cognitive agents in large urban areas considering sub-meter details of the environment. Details of the evacuation simulation software in the context of dynamical systems provide a common ground for the comparison with other evacuation simulation tools and a software specification language for further development. Examples of the implemented functionality for the creation of heterogeneous autonomous evacuees is further described. Different evacuees are modeled using different modes of evacuation (cars and pedestrians), complex sensing capabilities (visually perceiving the environment and identifying features, etc.) and cognitive capabilities (path planning considering the evacuees’ preferences and experiences, avoiding collisions with each other, etc.). Several demonstrative large urban area evacuation scenarios are presented in order to highlight the need for incorporating detailed models of the environment and the complex agent functions. These scenarios range from evacuations with different lighting conditions and environmental damages, to mixed mode evacuations using cars and pedestrians. Most of these scenarios, like night time evacuation in a damaged environment or mixed mode evacuation cannot be simulated with the simplified models without sacrificing accuracy.

中文翻译：

---

考虑详细模型的大规模疏散模拟：行为概况，环境影响和混合模式疏散

摘要本文提出了一种大规模疏散模拟器，该模拟器能够考虑大环境中的亚表细节，从而能够处理大城市中的复杂认知主体。动力系统环境中的疏散模拟软件的详细信息为与其他疏散模拟工具进行比较提供了共同基础，并为进一步开发提供了软件规范语言。进一步描述了用于创建异构自主撤离者的已实现功能的示例。使用不同的疏散模式（汽车和行人），复杂的感知能力（在视觉上感知环境和识别特征等）和认知能力（考虑避难者的偏好和经历的路径规划，避免彼此碰撞）对不同的避难者进行建模。等等。）。提出了几种示范性大城市疏散方案，以强调需要纳入详细的环境模型和复杂的代理功能。这些场景包括从具有不同照明条件和环境破坏的疏散到使用汽车和行人的混合模式疏散。在不牺牲精度的情况下，无法使用简化的模型来模拟大多数此类情况，例如在受损环境中的夜间疏散或混合模式疏散。