Pedestrians during evacuation exhibit complex and variable patterns of behavior, and understanding these patterns is extremely important for improving evacuation procedures and relevant regulations. The development of pedestrian dynamics for evacuation is an interesting and ongoing research topic in traffic science and engineering since the early 1990s [20]. In order to simulate the uncertainties and dynamics of pedestrians' behaviors, several famous models have been put forward in the domain of evacuation, such as cellular automaton model [21] and [22], agent-based model [23] and [24], and social force model [3]. Cellular automation is a discrete dynamic system with grid-based motion decision, which consists of a regular grid of cells, each in one of a finite number of states. The agent-based model simulates pedestrians with virtual agents and establishes social structures on a basis of a “bottom-up” order. However, it usually costs more computation time [25]. The social force model is in a continuous space and introduces the desired force to describe the inner drive of pedestrians to escape, especially under stressful situations [26].