“…The core idea of the former is to depict the passenger flow evolution in a subway network by simulating the passenger’s behavior [ 3 , 4 , 5 ], while the latter is to formulate an equivalent mathematical model by analyzing passenger route choice behavior based on travel cost [ 6 , 7 , 8 ]. Generally, these studies are based on the following assumptions: (1) each train has a fixed capacity [ 3 , 4 , 6 , 8 ]; (2) the passenger boarding process follows the FCFS (First-Come-First-Served) principle [ 6 , 7 ]; (3) ignoring the arrival time of an individual passenger [ 3 , 4 , 6 , 7 , 9 ]; and (4) neglecting the impact of network time-dependent state on passenger choice behavior [ 3 , 4 , 7 ]. However, the assumptions restrict the traditional methods in accurately depicting the factors influencing the passenger flow distribution [ 10 , 11 , 12 ], such as the in-train congestion [ 13 ], the passengers’ psychology during their travel process [ 14 ] and so on.…”