BayAreaMetro/travel-model-two

Updates to Intermediate Stop Location Choice Algorithm for CT-RAMP without TAPs

AshishKuls opened this issue · 1 comments

Existing implemented algorithm (with TAPs) to find the stop location alternatives on a Transit Tour evaluates:

For an Outbound Stop on Walk to Transit Tour:
Can you walk from Origin MGRA to Stop MGRA and then can you walk from Stop MGRA to Boarding TAP? or Can you walk from Alighting TAP to Stop MGRA and then can you walk from Stop MGRA to Destination MGRA? If yes, that stop MGRA is a potential stop location.

image

For an Outbound Stop on Drive to Transit Tour:
Can you Drive from Stop MGRA to Boarding TAP? Or Can you walk from Alighting TAP to Stop MGRA and then can you walk from Stop MGRA to Destination MGRA? If yes, that stop MGRA is a potential stop location.

image

There is similar logic's for Inbound Stop.

This algorithm needs to be revised in the CT-RAMP version without TAPs. Ideally, the logic to find stop location alternatives for transit tour would be to check - is there a transit path from origin to stop and is there a transit path from stop to destination. However, this would mean code will have to look up the transit skims from origin to all zones (alternatives) and from all zones to destination, for every stop for every tour, which is doable but computationally expensive.

We have currently implemented a simplified algorithm that is comparable to the existing algorithm and is computationally feasible.

For an Outbound Stop on Walk to Transit Tour:
If the Origin MGRA and Stop MGRA are in the same TAZ or if the Stop MGRA and Destination MGRA are in the same TAZ, then the Stop MGRA is a potential stop location (this works because we know that there is a transit path from Origin TAZ to Destination TAZ for TAZ-based transit skims).

image

For an Outbound Stop on Drive to Transit Tour:
If the Stop MGRA and Destination MGRA are in the same TAZ, then the Stop MGRA is a potential stop location. Note that the updated algorithm only allows the stop choice on the walk leg of the drive to transit tour. This is because we don't know which parking lot is going to get chosen for the drive leg and therefore making it impossible to identify a intermediate stop on the drive leg.

image

Team acknowledges that neither of existing or revised algorithms are correct way of finding the intermediate stop location choice and team will revisit this at some point and see if we would like to improve this in future.