Snake Game: A genetic neural network approach

There have been multiple attempts at playing the snake game with AI approaches which include Genetic Algorithm and Neural Network. This paper aims to research how tuning several parameters of the genetic algorithm and neural network will affect a snake agent in its performance. The parameters changed in this experiment are the mutation percent, percentage of best/worst performing, mutation intensity, and arena size to test the effects of each parameter on the performance of the snake agent. The consistency as well as performance of the snake agent are both observed closely in this study. We have found out that each parameter has its own degree of effect on the performance of the snake agent.