Rock Paper Scissors with genetic algorithm

branch main is the first attempt without much AI knowledge with human interface

branch genetic algorithm attempts to contain all children in 1 bot, follows the rule more but less effective

this branch runs all bots separately and modify the genes after match ends, more effective as bot1 will react differently to each individual

c and c2 is the same but separated the file to reduce human error switching between evaling and training
IO2.py and dllu.py are top 2 bots from rpscontest.com leaderboard

Usage

bot1 is the filename of the opponent
n is how many children you want in the population
-m x to specify x matches

start with randomized population

python genetic.py bot1.py c+n.py

continue with previous training

python genetic.py bot1.py c*n.py

evaluate result

manually copy the genes you want to c2.py
python rpsrunner2.py bot1.py c2.py