ray-dino / robby-genetic-algorithm

A genetic algorithm to create an evolving cleaning robot.

How to run

1. Install requirements.

pip install requirements.txt

2. Run the program.

python main.py

Project Outline

A genetic algorithm to create an optimal "Robby": a self-directed robot tasked to pick up cans littered around a 10x10 board.

Board

• A 10x10 matrix where each cell has a 50% of having a can.

DNA

• A sequence of 243 integers, each between 0-6.
• Dictates Robbie's behavior in any possible scenario.
• Can be created with random integers, or from a combination of 2 other DNAs with a small chance of mutation.

Robby

• A "robot" initially positioned at [0,0] of the board.
• Can determine the status of his surroundings.
• Has a DNA, that dictates behavior in any the possible scenarios.
• Has a lifespan of X turns, at the end of which a fitness score is calculated. +10 points for each can Robby successfully picks up. -5 if he crashes into a wall. -1 point if Robby tries to pick up a can when there isn't one.
• Can be born out of 2 other Robbies, with a new DNA that is a mix of the parents' DNAs, with a small chance of mutation.

Evolution

1. Start with P individuals (Robbies) that have random DNA. [Repeat the ff for G generations.]
2. Calculate the fitness for each individual in the current generation.
3. Apply evolution.

• Choose 2 Robbies as parents probabilistically based on fitness.
• "Mate" the 2 Robbies to create 2 children to add to a new population.

4. Once the new population is of size P, return to step 2.

To Do

• Concurrency
• Save the DNA of the fittest indvidual.
• Visualization showing a board and a Robby. Accepts input DNA object to determine the Robby's behavior.