"Explore the Technical Advisory Committee's meeting agenda covering updates, calibration, assignments, and more. Delve into the detailed 2024-25 work plan for GTAModel development, network maintenance, and staff allocations. Discover insights on road network calibration, NCS22 link functional class, and VDF calibration. Uncover the details regarding hyperparameters, assumptions, and more in the transportation management field." (498 characters)
Please find below an Image/Link to download the presentation.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author. If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.
You are allowed to download the files provided on this website for personal or commercial use, subject to the condition that they are used lawfully. All files are the property of their respective owners.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author.
Road VDF Calibration If we have ATC data at a variety of points, how to use these data in calibration? If we have speed data on a variety of links, how to use these data in calibration? ??? ? = ? ???????_????????? ?,?,??,? ??= ???? ?? ???? ? [1] ???????_?????????? = ???? ??? ?????????? ??????? ? ? = ?? ? = ?? ?? = ???? ? ? ?????? ?? ?? ???????? ? = ??????????? ??????? ?????????? hyperparameters ??= ?????? ???? ??? ?? ???? ? = ??? ??,??,??,??,????,?? [2] ??= ???????? ???? = ??? ? = ??,? ?,? = ???? ????? ? ??= ???? ????? ?? ??= ??.?? ????? ??? ??? ????= ???????? ?????,???? ? ??/ ? = ???? = ???????? ????? ??? ???? ????? ?,? ? ??= ?????????? ??? ???? ? ? = ?? ? = ??? ?? ???????? ???? ?????? ????? = ?? ,? ? ? = ??? ?? ???????? ???? ???? ?????? = ?? 6 ,? ?
Default assumptions: vdf = tangent function. I.e., not using, for example, Amin s functional form. ??,???? could be further parameterized as a function of other link attributes, adjacent land use, Not considering doing this at this point in time. Also, still assuming that links belong to exogenously defined function classes and don t have individually unique parameters. It is impossible to think of every link having unique parameters (over and above that this can t be done in Emme). This contracts with Amin s work, in which each link was individually parameterized using a machine learning algorithm and then he clustered links into classes. This requires a lot of consistent times series data to do, which are not available unless we purchase it from a data provider (Streetlight, HERE, etc.). Exogenous variables & parameters: ??,?, ??, ??,? ,? (Might also include screenline counts as additional targets?) Endogenous variables: ?,? Attributes, parameters to be estimated: ??, ????, ?? = ?? 7
Solution approach #1: Assign observed TTS flows by time period to the Emme network. Iteratively search for the parameter set that optimizes: 2+ + ?2 ? ? ?? ?? 2 min ?? ?1 ? ? ?? ?? ?1,?2 0,?????????? ???? ??,?1+ ?2= 1 Presumably use our Particle Swarm optimization procedure to solve for optimal values. Note that this requires running the Emme road assignment algorithm to evaluate each point examined in the solution space. Solution approach #2: Some sort of machine learning / clustering approach to learn the optimal values of ?? 8
Data Sources? MTO 2023 Travel Time Study: When available, will provide speeds on selected highway and arterial links, plus point-to-point travel times on routes surveyed. Includes measurement of HOV lane speeds vs. general lanes. 2020 survey results available for pre-COVID Jan-Feb 2020, but not clear how useful this is. Municipal data: ATC data? Speed data? If have sufficient data across road types, time of day, collection method(s) can be used. 9
Transit Assignment: Incorporating Reliability (1) Enlisting Prof. Shalaby & his research group (also NSERC PDF Dr. Mohammad Ansari Esfeh) to investigate the incorporation of transit reliability within the Emme transit assignment procedure. A challenging problem, since the Emme assignment procedure is static and reliability is inherently a dynamic concept. Two components: Schedule reliability: manifested in terms of variability in transit vehicle arrival times (variations of realized headways). Travel time reliability: variations in stop-to-stop travel times. 10
Incorporating Reliability (2): A Few Thoughts Wait time variability: Possibly replace the usual wait time factor (typically = 05. of headway) with a function of both mean headway & its variance. E.g.: E(wait time) = E(h)/2 + Variance(h)/(2*E(h)) h=headway Travel time variability: Add a penalty to the segment wait time (e.g., the standard deviation). Will need to be able to predict how future investments (e.g., dedicated transit lanes) change these reliability factors. E.g., factors will need to be a function of route/roadway design, traffic characteristics. Working on a research plan to investigate these ideas. 11
Prototype PTC Supply Model (1) We have developed a prototype PTC (Private Transportation Company: Uber/Lyft/etc.) supply performance model for a project with the City of Toronto Transportation Services Division to assess City PTC policies. Currently being implemented in XTMF. A preliminary upgrade of the GTAModel V4.2 tour-based mode choice model is also currently underway to interface with the new PTC supply model. Will be implemented in V5. 12
PTC Model (2): Preliminary Results: Simulating Jan. 2020 Performance Distribution of Driver Idle Time Durations Average Fare by Time of Day CO2 Emissions by Time of Day & Trip Stage Number of Drivers in Service by Time of Day 13
2024-25 Meeting Schedule No meetings in July & August! TMG Technical Advisory Committee Meetings (10:00-12:00): November 6, 2024 January 15, 2025 February 5, 2025 TMG Workshops (10:00-12:00): September 11, 2024 December 4, 2024 March 26, 2025 TMG Steering Committee Meetings (10:00-12:00): October 2, 2024 March 5, 2025 15
