Questions about Transportation Network Modeling and Analysis in R2D

[MohamedKhoudari]

Dear R2D Team,
I hope you are doing well.

I have been reviewing the recent updates to the transportation network example in R2D with the goal of using the platform for traffic simulation following an earthquake scenario. While going through the documentation and example files, I encountered a few concepts that were unclear and also had some questions specific to my application. I would greatly appreciate your clarification on the following points:

1. OD Matrix: Referring to the below figure, could you please explain the definitions of agent ID and person ID in the OD matrix? What is the distinction between the two?
   

2. Trip Assignment: When multiple trips originate from the same start node, and this node connects to multiple roads, what method does R2D use to distribute the demand across the network? Is it based on shortest path, user equilibrium, gravity model, …?

3. Road Capacity Calculation: How is the capacity of each road determined within R2D? What variables are used in the calculation (e.g., speed, number of lanes, …)?

4. Variable Definitions: Referring to Figure 2, could you clarify the meanings of the variables veh_current, vol_true, and vol_tot?
   

5. Incorporating Debris Effects on Roads: I understand that R2D accounts for transportation network damage primarily from ground deformations and adjusts road capacities based on the calculated damage state and a predefined capacity map. I would like to consider the additional impact of building debris obstructing roads. My idea is to modify the damage state of roads based on an external debris footprint analysis.
   However, I noticed that R2D proceeds directly from the damage state analysis to the traffic simulation, without providing access to the intermediate damage state output files. Is there a way to pause the workflow after the damage state is computed, so I can modify these files before the traffic simulation step? Alternatively, is there a method to input modified damage state files directly into the traffic simulation module?

Thank you very much for your time and support. I look forward to your guidance.

Best regards,
Mohamed Khoudari

[jinyan1214]

Hi Mohamed,

Thank you very much for your questions and your interest in using this tool. Below are some responses:

1. The agent ID is an index that is used in the residual demand simulation. This OD matrix is a subset of one of the travel demand models that the Bay Area MTC produced. And the person ID is the index used in the MTC OD matrix. Using the person ID, you can find more information from the MTC OD matrix, including the category (e.g., Mandatory/Nonmandatory) and tour_purpose (e.g., school, work, social, etc)
2. The current implementation of R2D uses a shortest path-based assignment approach to distribute traffic demand across the network. Specifically, the model leverages the Pandana package (https://github.com/UDST/pandana/tree/dev) to calculate optimal paths. The default edge weight for the shortest path algorithm in R2D is the estimated travel time, dynamically updated at each simulation interval using the Bureau of Public Roads (BPR) function, which reflects congestion effects.
3. In the normal (pre-disaster) state, road capacities are calculated based on the number of lanes, using a default value of 1800 vehicles per lane. Thus, the total capacity of each road is determined by multiplying its number of lanes by 1800 vehicles per lane. After an earthquake, the capacities of damaged roads are calculated using a user-defined capacity map. This map allows users to specify different capacity reduction ratios corresponding to various road damage states. In Example 14 of R2D, a road in damage state 1 retains 97.5% (ratio = 0.975) of its original capacity, whereas a road in damage state 2 only retains 25% (ratio = 0.25). The relationship in Example 14 is a demonstration of how to define the capacity reduction, and it was estimated according to a publication: Liu, K., Zhai, C., Dong, Y., & Meng, X. (2023). Post-earthquake functionality assessment of urban road network considering emergency response. Journal of Earthquake Engineering, 27(9), 2406-2431.
4. "veh_current" indicates the current number of vehicles present on a given road segment during the simulation. It is continually updated based on incoming and outgoing vehicle flows. "vol_true" represents the actual observed traffic volume recorded for each road segment within the current simulation interval (e.g., a quarter-hour). "vol_tot" refers to the total cumulative volume on a road segment within the simulation period or analysis interval.
5. This can be achieved. At the top of the R2D software, there is a "Tools" menu, and a "Residual Demand" tool can be found in the "Tools" menu. This toolbox allows one to run a traffic simulator using existing R2D analysis results. You can potentially edit the R2D results file in your "Local or Remote WorkDir/tmp.SimCenter/Results" folder to include debris-induced roadway closure. Then, you can run a traffic simulation by providing the same input files as in the main R2D window and clicking "Run Simulation" to simulate traffic using the modified results file. A screenshot of the "Residual Demand" tool is attached below. If you need more flexible behavior, it can be achieved by interacting with our backend simulation code. Please let us know if this is needed and we can schedule meetings to discuss further.

I hope the above is clear.