Setup - RSGInc/SeaCast GitHub Wiki
Scenario Input Files
The SeaCast Scenario Inputs contains scenario inputs to the SeaCast model. For the setup purposes we will assume that the respository is downloaded to C:\users\jane\seacast_test\seacastscenarioinputs-master folder.
Inputs are made available in hierarchy by year for land use and network inputs. This allows the user to specify various alternatives for each year, and allows a combination of different land use and network combinations. In the provided 2018 input data the directory structure is as follows:
- v3.0.0
- base_year
- 2018
- all base year inputs for 2018.
- 2018
- db
- this single database file contains data for multiple years. Data is filtered based on the specified model year
- landuse
- 2018
- land_use_2018
- all land use data corresponding to year 2018 and the base v3.0 policies
- land_use_2018
- 2018
- networks
- 2018
- networks_2018
- all network data corresponding to year 2018 and the RTP assumptions
- networks_2018
- 2018
- base_year
Input Configuration
The primary control configuration for typical runs is the input_configuration.py file in the root directory of the SeaCast repository. This file should be updated to identify the location of model inputs prior to running.
Input Paths
Near the top of input_configuration.py, ensure that following parameters are accurate and point to the proper input set.
Update the soundcast_inputs_dir with the path (with forward slashes) where the SeaCast scenario inputs were downloaded.
model_year = '2018'
base_year = '2018'
landuse_inputs = 'land_use_2018'
network_inputs = 'networks_2018'
soundcast_inputs_dir = 'C:/users/jane/seacast_test/seacastscenarioinputs-master/inputs'
Model Component Control
The default setting for SeaCast is to run all model components, including
- freeflow assignment and skimming
- Daysim choice models for all simulated person-days
- global iterations between assignment and choice modeling until convergence
- truck, bike, and external trip assignment
- summaries
These components may be run or skipped individually, as defined by parameters in input_configuration.py. The default settings should not be changed to run a standard full model run. If a run is interrupted or only assignment or summary scripts need to be run, True/False statements can be used to control whether model components are executed. The following table describes the configurable parameters in input_configuration.py.
| Variable | Default Setting | Description |
|---|---|---|
| model_year | '2018' | Model year used to look up year-specific inputs |
| base_year | '2018' | Should be always be 2018 regardless of scenario, unless using another base year |
| soundcast_inputs_dir | 'C:/users/jane/seacast_test/seacastscenarioinputs-master/inputs' | top-level location of input directory |
| landuse_inputs | 'land_use_2018' | folder name for land use inputs nested within scenario inputs directory |
| network_inputs | 'networks_2018' | folder name for network inputs nested within scenario inputs directory |
| run_accessibility_calcs | True | Calculate land-use and network accessibility measures from all-streets network; only needs to be performed once during model run |
| run_setup_emme_project_folders | True | Create empty (or overwrite existing) Emme projects at initial run setup |
| run_setup_emme_bank_folders | True | Create empty emmebanks at initial setup |
| run_copy_scenario_inputs | True | Copy input files to local SeaCast directory |
| run_import_networks | True | Load networks to each time-of-day Emme project |
| run_skims_and_paths_free_flow | True | Default for SeaCast is free-flow assignment |
| run_skims_and_paths | True | Majority of trip assignment and skim building occurs when True |
| run_truck_model | True | Generate truck trips and add to network |
| run_supplemental_trips | True | Separately consider external and special purpose trips not considered by Daysim |
| run_daysim | True | Run the main activity-generation portion of the model |
| run_summaries | True | Run a set of network and activity summaries to produce CSV and HTML outputs |
| include_av | False | Model inclusion of automated vehicles |
| include_tnc | True | Model including of TNC/ridehailing vehicles |
| tnc_av | False | Allow TNCs to operate as AVs |
| include_tnc_to_transt | False | AV drop offs to transit |
| include_knr_to_Transit | False | Allow Kiss and Ride drop-offs at transit stations |
| include_delivery | False | Allow for additional delivery trips to neighborhoods |
| add_distance_pricing | False (in base year; True in future years generally) | Apply per-mile pricing |
| distance_rate_dict | {'am' : 13.5, 'md' : 8.5, 'pm' : 13.5, 'ev' : 8.5, 'ni' : 8.5} | Prices per mile applied when |
| num_enplanements | 24894338 | Number of enplanements at SeaTac airport (should be updated for future year) |
| connecting | 7184927 | Number of connections at SeaTac airport (should be updated for future year) |
| households_persons_file | r'inputs\scenario\landuse\hh_and_persons.h5' | synthetic population file to be used for household resampling |
| sampling_option | 2 | one of the three sampling strategy used as described in pop_sample_district below |
| pop_sample_district | {'City of SeaTac':[1,4,2], 'Rest of King County':[1,1,0.75], 'Rest':[1,1,0.75],} | sampling strategies available for household resampling |
| zone_district_file | r'inputs\model\lookup\hh_sampling_region_taz.csv' | intermediate input file for household resampling |
| taz_sample_rate_file | r'inputs\model\lookup\taz_sample_rate.txt' | intermediate input file for household resampling |
| run_integrated | False | For integration with urbansim |
| should_build_shadow_price | False | Drive-to-transit shadow pricing control |
| delete_banks | False | Method to minimize run size by removing large Emmebank data files |