deapsleep 1.0.0

DeapSleep is a DEAP-based evolutionary computation toolkit for testing dropout in genetic algorithms

DeapSleep

DeapSleep is a DEAP-based evolutionary computation toolkit for testing dropout in genetic algorithms (GAs). Random deactivation of individuals or decision variables might serve as diversity-induction mechanism, improving the algorithm’s ability to navigate rugged fitness landscapes with multiple peaks and valleys. The toolkit provides a GUI for easily optimize, plot and compare the results of optimization using a vanilla or dropout-based GA. DeapSleep is fully Dockerized, but using Docker is optional, as it can also be run directly from the command line.

DeapSleep includes the following features:

• Genetic Algorithm using DEAP implementation
• Single and multi-objective optimization (NSGA-II, NSGA-III)
• Hall of Fame (Hof) of the best found individuals
• Pymoo benchmark functions and corresponding internal .yaml configurations
• An interactive GUI for:
  • optimizing benchmark functions or custom problems using
    • vanilla GAs
    • individual dropout (IDrop), i.e. decision variables deactivation
    • population dropout (PDrop), i.e. complete individuals deactivation
    • both (I&PDrop)
  • plotting converge-to-target results and HoF statistics over n runs
  • comparing vanilla GAs vs dropout versions

Disclaimer: at the moment, DeapSleep does not allow to use custom problems.

Installation

1. Docker installation (necessary for using the GUI)

Clone the repository

git clone https://github.com/dganci/deapsleep.git
cd deapsleep

Build the image

docker build -t deapsleep .

Run the container (and save the results in <YOUR_DIRECTORY>

docker run -it --rm -e DISPLAY=$DISPLAY -v /tmp/.X11-unix:/tmp/.X11-unix -v "<YOUR_DIRECTORY>:/app/results" deapsleep

2. Python installation

Clone the repository

git clone https://github.com/dganci/deapsleep.git
cd deapsleep

We recommend to use a virtual environment, e.g.

python -m venv venv
# Linux/macOS
source venv/bin/activate

Then

pip install -r requirements.txt

Examples of usage:

• for optimization

python3 -m deapsleep.main.optimize --config single.ackley -i --version='baseline_example' --n_runs=10 --ngen=200

python3 -m deapsleep.main.optimize --config single.rastrigin -i --version='I&PDrop_example' --n_runs=30 --ngen=1000 --n_var=10 --popD_rate=0.7 --indD_rate=0.2

• for plotting

python3 -m deapsleep.main.plot --problem griewank --version='baseline_example' --dirname="${PWD}/results"

• for comparison

python3 -m deapsleep.main.compare --config multi.zdt1 -i --version1='baseline_example' --version2='IDrop_example'