Analisis Reinforcement Learning pada Game Box Endless Runner 2D

Abstract

This study aims to compare the effectiveness of three reinforcement learning algorithms, namely DQN, A2C, and PPO, in optimizing the performance of an agent in the 2D Box Endless Runner game. The study evaluates the performance of each algorithm by examining the rewards generated during testing episodes, as well as the generalization of the trained models to different game environments. The experimental results show that DQN, A2C and PPO produce effective models for the Endless Box Runner 2D game problem. However, this result should be considered as a general indication and cannot be directly generalized to other problems. Furthermore, the trained models in the same initial environment but different algorithms show good learning abilities in the generalization test.