Resumen
Numerous pedestrians interact with the subway station space by finding entrances into this closed area to use the subway system; further, they may use transfer transportation facilities or the complex functions nearby, such as commercial. Many studies examine pedestrian behaviors in subway stations, but most focus on special situations such as disasters and evacuation. Because it is important to analyze gait patterns in everyday situations, this study aims to verify the explanatory power of actual gait behavior by using space syntax theory in constructing an optimal agent-based model. To this end, first, pedestrian characteristics and space types are classified using pedestrian data from Gangnam Station. Second, the depthmapX program is used to develop an appropriate agent-based model for stations. Third, a simulation is run to calculate the frequency of the agent movement at each gate, which is matched with the observed pedestrian volume. Fourth, the relationship between the frequency of the agent movement and pedestrian volume is analyzed using Statistical Package for the Social Sciences. The results show that although agent-based models have limitations in explaining pedestrian patterns in the entire subway station, they are capable of explaining these patterns along the shortest paths between ticket gates and station entrances.