Program Overview

This program generates a population of Variational Quantum Circuits and optimizes them using an Evolutionary Algorithm. Each circuit is evaluated in the Coin Game environment, where agents play a round of 50 steps against themselves. The agents' fitness is assessed based on their score in the Coin Game and the size of their circuits.

Output: The program generates a .csv file in the data directory, containing information from all generations, including performance metrics from the Coin Game and circuit sizes.

Optional:

• create virtual environment (in root of project): python3 -m venv venv
• activate virtual environment: source venv/bin/activate

Running the project:

• install dependecies: pip install -r requirements.txt
• run main.py with arguments : python src/main.py
• mandatory arguments:
  • seed (--seed / -s)
    • with some numerical seed value
  • architecture concept (--type / -t)
    • Layer-Based: layer,
    • Gate-Based: gate,
    • Prototype-Based: prototype INFO: the size specifications for all concepts are in the parse_args function
  • evolution type (--evolution / -e)
    • mutation only: mut-only
    • recombination and mutation: mut-recomb
  • number for architecture mutations (--num-arch-mut / -a)
    • integer value
    • set this to specify a number of architecture mutations
• optional arguments:
  • dynamic mutation power (--dynamic-mut-pow / -d)
    • float between 0 and 1
    • if set dynamic mut pow in main.py is used, else the static mut pow in main.py is used
  • increased exploitation in last quarter of the generations (--exploit / -x)
    • no value
    • if set selection changes to Truncation Selection with top 5 agents in last quarter

---