Combination of Genetic Algorithm and Neural Network for Independent Game Playing

In this paper, in order to tackle the challenge of a game playing agent, a combination of algorithms with the aim to play and win the game was introduced. The agent will learn to play a simplified version of a very popular game Super Mario. Even though launched almost four decades ago, it is considered as one of the most famous game, played until now by humans through years with many competitions also organized. Researchers argued that it is quite a hard game to win by humans. Our approach is to use genetic algorithm with the combination of neural networks to achieve the aim. The experiments conducted show promising results by reaching the requested score for the provided time.