BayAreaMetro / tm2py

Travel Model Two - python edition
https://bayareametro.github.io/tm2py/
Apache License 2.0
2 stars 8 forks source link

🚀 Feature: Test OpenPaths EMME 2024's Ignore Cosmetic Nodes Feature in Highway Assignment #151

Closed i-am-sijia closed 1 month ago

i-am-sijia commented 3 months ago

User Story

The existing TM2 roadway network is highly detailed, including every street, bicycle path, and sidewalk. When the roadway network is extracted from this network, it contains numerous nodes that are unnecessary for roadway assignment.

OpenPaths EMME 2024 allows ignoring the cosmetic nodes in the highway assignment to speed up run time. We will test it to confirm the details.

Progress:

Priority

High

Level of Effort

Low

Resolution Ideas

With the new EMME, we should be able to add an assignment argument to trigger the ignoring of cosmetic nodes.

Project

Task Order 5

Who should be involved?

Users: @gregerhardt, @FlaviaTsang, @lmz Reviewers: @i-am-sijia, @DavidOry

Risk

We need to confirm if the output assigned network is in the original format with the cosmetic nodes, or in a simplified format that does not have the cosmetic nodes. It sounds like it should be the former.(?)

Tests

i-am-sijia commented 3 months ago

OpenPaths testing plan:

lachlan-git commented 1 month ago

Step 1,2 and 3 have been tested. Note Emme bank contained links FT 1-7:

Run Number Desciption VDI Machine number of iterations [EA, AM, MD, PM, EV] time for for highway assignment and skims
23 Base case
FT 7 included included
Network Accelerate = False
Lachlan Tier 2 Pool B [3, 13, 10, 30, 4] 6 hours and 15 minutes
24 same as base Case except:
network accelerate = True
Lachlan Tier 2 pool B [3, 12, 10, 24, 4] 4 hours and 15 minutes
25 same as base Case except:
network accelerate = True
Emme remove_cosmetic_nodes
Lachlan Tier 2 pool A [3, 12, 9, 23, 4] 3 hours and 30 minutes
lachlan-git commented 1 month ago

To Perform QA on the above results, All volumes of all links are compared against run 23 (the base run above). Slope and R2 are reported for each func type and area compared against.

Below table is "link_wise" comparisons, note run 24 will have some links dropped due to removing aesthetic nodes.

Base Run Comparison Run Facility type Slope R Value
0 emme Open Paths emme 4.6.1 Freeway 0.94 1
1 emme Open Paths emme 4.6.1 Expressway 0.95 1
2 emme Open Paths emme 4.6.1 Ramp 0.94 0.98
3 emme Open Paths emme 4.6.1 Divided Arterial 0.93 0.99
4 emme Open Paths emme 4.6.1 Undivided Arterial 0.94 0.99
5 emme Open Paths emme 4.6.1 Collector 0.95 0.99
6 emme Open Paths emme Open Paths Network Accelerate Freeway 0.98 1
7 emme Open Paths emme Open Paths Network Accelerate Expressway 0.98 1
8 emme Open Paths emme Open Paths Network Accelerate Ramp 0.97 0.99
9 emme Open Paths emme Open Paths Network Accelerate Divided Arterial 0.97 1
10 emme Open Paths emme Open Paths Network Accelerate Undivided Arterial 0.98 1
11 emme Open Paths emme Open Paths Network Accelerate Collector 0.98 0.99
12 emme Open Paths remove Cosmetic Nodes Freeway 1 1
13 emme Open Paths remove Cosmetic Nodes Expressway 1 1
14 emme Open Paths remove Cosmetic Nodes Ramp 0.99 0.99
15 emme Open Paths remove Cosmetic Nodes Divided Arterial 0.98 0.99
16 emme Open Paths remove Cosmetic Nodes Undivided Arterial 0.99 0.99
17 emme Open Paths remove Cosmetic Nodes Collector 0.98 0.98
DavidOry commented 1 month ago

@lachlan-git: can you please edit the above table to:

  1. Replace the "runXX" code with a more meaningful description.
  2. Replace the "func_types" code with the word labels.
  3. Use two decimals for the quantities.

thanks.

lachlan-git commented 1 month ago

done