The core functionality of the price simulator is in folder price_simulator/src/algorithm
.
The components of the simulator are programmed as separated objects.
Thus, agents, demand, environment and policies have own files and classes.
This object-oriented approach allows a free combination of different agents, demand models, or policies.
New agents or more sophisticated demand models integrate straightforwardly.
main.py
shows an examplary entry point to start the simulation.
The environment is able to handle an arbitrary amount of different agents.
This makes a huge variety of experiments easily possible.
To start, you'll need docker
installed. When done, run this line in price_simulator dir:
docker build -t price-simulator:1.0 .
It will build docker image, used for running other commands. You can execute scripts using Docker. Docker must be build to reflect changes.
docker run price-simulator:1.0