Optimal-Strategy-Transit-Assignment is a repository for the static (frequency-based) transit assignment python code. The model was proposed by Spiess and Florian (1987).
This is an assignment model for the transit systems where the service is perceived in terms of frequencies. The passengers traveling between an origin-destination pair served by multiple transit lines are faced with the problem of which transit route to board. Due to the randomness in the arrival time of buses, the passengers adopt strategy which minimizes the expected travel time. A strategy is defined as a set of rules when applied helps a passenger to move from origin to destination in a transit network. Nguyen and Pallottino 1988 formalized the idea of strategy in a graph-theoretic framework and called it a hyperpath. Let v_ak be the flow on link a destined to destination k and let W_ik be the wait time of all the passengers visiting node i destined to destination k. The, the transit assignment can be formulated as a linear program given below:
To avoid solving the large LP, Spiess and Florian proposed a primal-dual algorithm which is coded in python.
For the assignment, you need three types of data:
- Network data: For this, you can use R file named
DataPrep.R. Change the location of the GTFS ZIP file in the script and run it. This will create nodes and links for the network. - Zones data/Access links data: You need data related to traffic analysis zones from where the demand is generated or destined to. I will leave this up to you to create access_links from zones to transit stops and vice versa. Check the data format
ft_input_accessLinks.dat - Demand: You need demand between various zones.
The code is under progress. Please do not use this code.