An algorithm for transit disruption mitigation - strategy seletion phase (resource allocation)1
Run main.py [Q_scenario] [q_0] [q_max] [T_scenario] [alpha]
@args Q_scenario: the name of demand pattern
@args q_0, q_max: specification the level of demand, like "uniform", "convex", "concave", "increasing", "decreasing"
@args T_scenario: the name of the disruption distribution, like "uniform", "exponential", "normal", "Dirac_0", "Dirac_Tub", "bi_Dirac"
demand pattern illustration:
disruption distribution illustration:
The evaluation module takes time-dependent demand, disruption distribution, and mitigation plans as input, and outputs the user and operator cost in the horizon. Time is discretized into one minute intervals when accumulating the user costs. The user demands in each one-minute interval are assigned according to the capacity constraints at that time. Not enough capacity on the shortest path means that users will detour to longer distance paths. User wait cost depends on average headway.
User cost computation:
Qi Liu(ql375@nyu.edu), Joseph Chow(joseph.chow@nyu.edu)
Footnotes
-
some of the code is designed specificaly for example 1; changes are needed if you want to apply it to other networks! ↩