lighthouse

This is a light-weight, simplified ABM.

Usage

This project uses uv as Python manager. To install uv, please visit https://docs.astral.sh/uv/getting-started/installation/

Once uv is installed on the machine, create a new Python environment for lighthouse and install dependencies.

uv sync --locked

If running multi-processor, then need to disable multi-threading (once per session)

$env:MKL_NUM_THREADS = "1"
$env:OMP_NUM_THREADS = "1"
$env:OPENBLAS_NUM_THREADS = "1"
$env:NUMBA_NUM_THREADS = "1"

To run the model with test data, use the following command:

cd model
uv run --project .. activitysim run -c configs_mp -c configs -d data -o ../output-24-2K --ext extensions --households_sample_size 2000
# run all households in the test data
# uv run --project .. activitysim run -c configs_mp -c configs -d data -o output --ext extensions

Monitoring

See Monitoring a model run for CPU/memory sampling and progress tracking from ActivitySim run plans and logs.

Constraint components

See Fixed work schedule for the industry/income logit prototype, configuration, outputs, and validation. Actual work-hour scheduling is deferred.

Contents

• model: ActivitySim inputs (configs, data) for the lighthouse model. Currently model/data contains test-scale data from CTPS.
• notebooks: Demo notebooks to test if the model still works.
• src/lighthouse: Python code used to implement this model. This may grow to include extensions to ActivitySim for things we want the lighthouse model to do.

Development

To point to a local repo, add this to the pyproject.toml file with path to your ActivitySim repo [tool.uv.sources] activitysim = { path = "C:/projects/activitysim", editable = true }