Skip to content

Installation

Requirements

Software Version Used by
Python 3.10 or later every step
Java (JDK) 21 SILO, MAST, MATSim
Maven 3.9 or later building the Java engines
Memory 32 GB or more SILO (24–28 GB heap), MAST (13 GB), MATSim (32 GB)

1. Get the code and the miles command

git clone https://github.com/Tomal121186621/MILES.git
cd MILES
python -m venv .venv && source .venv/bin/activate
pip install -e ".[population,docs]"
miles --help

pip install -e . installs the miles command from your clone, so it always runs the code you have checked out.

2. Configure

cp miles.example.toml miles.toml

Set paths.workspace to the folder that holds the data and receives the runs, and the locations of the external checkouts (ABIT resources, MATSim inputs). Then check:

miles config

3. Estimation environment

python -m venv ~/.venvs/biogeme
~/.venvs/biogeme/bin/pip install biogeme==3.3.5 pandas pyarrow matplotlib ipykernel nbformat
~/.venvs/biogeme/bin/python -m ipykernel install --user --name biogeme --display-name "Python (biogeme)"

Set python.biogeme in miles.toml to this interpreter.

4. The Java engines

git clone https://github.com/msmobility/silo.git && cd silo
git checkout 3066ca9
git apply ../MILES/landuse/engine/silo_changes.patch
cp -R ../MILES/landuse/engine/src/* .
mvn compile -pl siloCore,extensions/schools,useCases/maryland
cd activity/engine/ABIT
mvn compile                     # main class abm.RunAbitMarylandLos

The engine reads its zone resources from the ABIT checkout set as paths.abit_home.

cd assignment/engine/matsim-run
mvn package                     # → target/baltimore-matsim-1.0.jar (MATSim 2024.0)

5. Data

MILES ships no data. Download the public sources and request the restricted ones as listed in Input data, into the workspace. After the land-use coefficients are written, copy the engine input files into the workspace with:

miles sync