marketdesign 0.0.1a1

This project is a python implementation of the market mechanism described in the paper 'Shapley value based pricing for auctions and exchanges' (Lindsay 2018).

Market Design

This project is a python implementation of the market mechanism described in the paper "Shapley value based pricing for auctions and exchanges" (Lindsay 2018). The project uses a bidding language that extends the version in the paper to allow bidders to express their preferences more concisely.

Installing on Windows

Install git and Python 3.x if not already installed.

# Get a copy of the project. 
git clone https://github.com/luke777/marketdesign.git
# Set up and activate the virtual environment 
cd marketdesign
py -m venv env
env\Scripts\activate
py -m pip install -r requirements.txt
# Verify that the command-line solver runs
py solver.py examples\Lindsay2018_t1_seller_and_2_buyers.txt

If Windows PowerShell gives the error message "running scripts is disabled on this system", try executing the following Set-ExecutionPolicy Unrestricted -Scope Process

On Linux/macOS try using python3 instead of py to run python and source env/bin/activate instead of .\env\Scripts\activate to activate the virtual environment.

Command-line solver

The program solver.py reads in bids from a file and outputs the winning bids and payments. If the python virtual environment is not already activated, activate is as follows.
env\Scripts\activate
The solver can then be run.
py solver.py examples\Lindsay2018_t1_seller_and_2_buyers.txt

Three file formats are supported. The directory examples contains examples of each of the formats. The cases are taken from Lindsay 2018.

• Text This format is the most concise. Bids are specified in a text file with one line per bidder.

• CSV This format is convenient if bids are prepared in a spreadsheet.

• JSON This format is relatively more verbose but allows bids to be easily prepared and analyzed using most programming languages.

By default, the results of solving the market are shown onscreen. The output can be directed to a file as follows.
py solver.py examples\Lindsay2018_t2.1_two_demand_types.csv -o output.json
The format of the output is determined by the filename extension. As with inputs, the supported formats are text (.txt), csv, and json.

Web solver

The program web-solver.py starts a micro web service that allows the market solver to be accessed online.

If the python virtual environment is not already activated, activate it with env\Scripts\activate
Then start the service with
py web-solver.py

Opening http://localhost:5000 in a browser should display links to the following.

• A form for submitting bids in text format.
• A form for uploading a file with bids in cvs format.
• A form for submitting bids in json format.
• An example of a graphical interface that lets users experiment with different bids. The bids are captured using javascript and sent to the server in json format.