Solving stochastic frequency-based transit assignment using the Gammit model is an approach designed to model how passengers choose transit paths (hyperpaths) in networks where they face uncertainty regarding service, such as bus arrival times at stops, according to sciencedirect.com and ideas.repec.org. This method accounts for both pre-trip and en-route path choices and handles overlapping paths better than traditional models.
Here are key aspects of this method, based on research into this topic:
Gammit Model Functionality: The Gammit model uses a specific formulation for link costs, represented as
. In this formulation, β = τ is common to all arcs, and the dispersion parameter τ is often assumed to be 0.2.
Hyperpath Modeling: The approach models travel strategies using hyperpaths (sets of attractive lines) to represent the stochastic nature of user choices, which is particularly relevant for urban transit networks where travelers may not have complete information.
Methodology and Algorithm: The problem is solved using Monte Carlo simulation techniques, often comparing Sobol numbers with Mersenne Twister methods. These techniques are embedded within MSA-based (Method of Successive Averages) algorithms to achieve equilibrium assignment.
Transit Network Features: The model considers both in-vehicle links (segments of transit lines) and waiting links (connections between boarding and alighting nodes) within the transit network graph G(N,L).
Advantages: This approach properly models the effects of hyperpath overlapping, which is a common limitation in other transit assignment models. For more context, I can:
Find comparisons between this and other methods (like Probit or Logit).
Search for specific case study results from the referenced papers. Detail the “MSA-based algorithms” mentioned.
Leave a Reply