Pedestrian Simulation Considering Stochastic Route Choice and Multidirectional Flow

This paper proposes a framework of a dynamic pedestrian model integrated with a strategic route choice model and multi-directional flow propagation. Despite a vast number of proposed pedestrian models especially in the last decade, very few of these considered both pedestrian’s route choice behavior and physical movement in a single integrated framework. In an open space, even after the discretization of the space, a massive number of possible paths for pedestrians can possibly be defined. However, psychologically we obviously do not consider all these paths as distinct options. Thus, pedestrians in this model are assumed to choose a series of consecutive sub-areas to traverse along at first instance. Then, the model allocates the actual flows on each detailed path based up on an assumption of dynamic user optimum (DUO). Combining these two stages of flow allocation, the model thus assumes a hierarchical decision of pedestrian’s route choice. In this paper, a modified celltransmission model (CTM) is adopted to represent the physical phenomena of congestion and dynamical propagation of pedestrian flows. A key difference to the original CTM is that the pedestrian flow is multi-directional by its nature. Thus, the paper proposes an approach to extend the analysis in the original CTM to consider the multi-directional movement. The paper also illustrates the plausibility of the proposed modified CTM by testing with a test case.