Dwell Time vs. Bus Stop

Peter,
The folks there now can comment about how it is currently done, but when I
was at SFCTA we estimated dwell time functions from APC data and added them
to the cube line files and did iteratively assigned them using TRNBUILD
using MSA to adjust the dwell times with each iteration until the travel
time skims converged (note - in the final global model iteration we looked
for convergence of the transit route volumes). The scripts for reading the
assignment files and adjusting the dwell times in the tp+ files were
written in Python and were part of SF-CHAMP.

The process is documented in this paper:
https://link.springer.com/article/10.1007/s11116-012-9405-x
...but I"m sure the folks at SFCTA can give you a tour.

cc: LIsa Zorn and Dan Wu who actually did most of this work in case they
have anything to add

Best,
Elizabeth Sall

UrbanLabs LLC

On Wed, May 10, 2023 at 4:01 AM Peter Chen wrote:

> Dear Colleagues,
>
> Our agency is trying to develop different bus transit scenarios in order to
> attract more passengers, either via increasing frequency or decreasing
> dwell time. One of the scenarios to decrease dwell time is to decrease the
> number of bus stops. However, it is a trade-off between ridership increase
> (less dwell time) and decrease (less accessibility). We want to test which
> works better. Our model is based on Cube TP+. Does anyone have experience
> to simulate the bus dwell time in TP+ (TRNBUILD)? We know PT has a dwell
> time statement; unfortunately our model uses TP+, not PT. Really appreciate
> any feedback!
>
> Peter Chen, Santa Clara Valley Transportation Authority
> --
> Full post: https://tmip.org/content/dwell-time-vs-bus-stop
> Manage my subscriptions: https://tmip.org/mailinglist
> Stop emails for this post: https://tmip.org/mailinglist/unsubscribe/13963
>

Hi Peter
I am not providing a Cube technical response (unfortunately!) and hopefully someone else will.

However my initial thought with your requirement is that (in general) caution needs to be applied when reviewing individual transit route performance within a regional model and this is even more the case for individual bus stops.

With regards to removing bus stops, it might be easier to go back to ‘first principles’ and base the analysis on current ridership and route configurations rather than extensive modelling. See below example of work TransLink did in Vancouver:

Bus Stop Balancing | TransLink

Full disclosure, I was NOT involved in that study and provided for reference.

Regards,

Dan Gomez-Duran
Director
K I Squared
Web: ki-2.com
Tel: +1 778 951 6198

From: chpchen=gmail.com@mg.tmip.org > On Behalf Of Peter Chen
Sent: Wednesday, May 10, 2023 4:02 AM
To: TMIP >
Subject: [TMIP] Dwell Time vs. Bus Stop

Dear Colleagues,

Our agency is trying to develop different bus transit scenarios in order to
attract more passengers, either via increasing frequency or decreasing
dwell time. One of the scenarios to decrease dwell time is to decrease the
number of bus stops. However, it is a trade-off between ridership increase
(less dwell time) and decrease (less accessibility). We want to test which
works better. Our model is based on Cube TP+. Does anyone have experience
to simulate the bus dwell time in TP+ (TRNBUILD)? We know PT has a dwell
time statement; unfortunately our model uses TP+, not PT. Really appreciate
any feedback!

Peter Chen, Santa Clara Valley Transportation Authority
--
Full post: https://tmip.org/content/dwell-time-vs-bus-stop
Manage my subscriptions: https://tmip.org/mailinglist
Stop emails for this post: https://tmip.org/mailinglist/unsubscribe/13963