Agent-based model using MESA package modeling a two-sided market, pay-as-clear pricing rule based micro-grid energy market. Goal is to simulate a grid with households using plug-and-play bidding strategies and devices according a predefined grid set-up. Includes 3 pricing algorithms; Pay as Bid (i.e. discriminatory auction), Pay as Clear (i.e. uniform auction) and the McAfee Auction.

This is a project in Python 3.6.3. We advice to use a virtual environment.

Clone repo into a dedicated folder (in your home directory). In terminal, located inside this folder, do:

git clone https://github.com/dirkbig/pac

Install dependencies (MESA, matlibplot, seaborn, and some standard ones), by directing pip-installer towards the requirements.txt

pip install -r requirements.txt

Run the micro-grid simulator file in the command-line

python run_microgrid.py

Since dependency on Matlibplot, the 'python is not installed as framework' error could be occurring. If so:

echo "backend: TkAgg" >> ~/.matplotlib/matplotlibrc
python run_microgrid.py

A plot is automatically generated and opened, showing all bids and offers in a aggregated demand/supply curve, with an market clearing quantity. For pay-as-clear, this corresponds directly with the clearing price on the Y-axis. Not adviced when running large amounts of steps.

When running data,py as main, agent profiles are plotted, according to the grid_config.py file. grid_config.py is used by microgrid simulation.

This is a research collaboration between Grid Singularity / Energy Web Foundation - Reiner Lemoine Institut

• Dirk van den Biggelaar - GSy
• Fatuma Mohammed Ali - GSy
• Marlon Fleck - RLI
• Jorn Hartmann - RLI