Authors : Simran Bhojwani, Sagar Ganiga, Harsh Jain, Shravan Jain, Riya Karia, Nausheen Sayed, Jayashree Hajgude.
Volume/Issue : Volume 4 - 2019, Issue 4 - April
Google Scholar : https://goo.gl/DF9R4u
Scribd : https://bit.ly/2QTbjdB
This paper describes the implementation of Reinforcement Learning algorithms on FPS Games by applying AI Agents. The steps undertaken to build an FPS Game using various game engines and reinforcement learning are outlined. The game consists of player AI and opponent AI competing with each other in which the goal is to kill the opponent. The algorithm has been tested on different maps of vizDoom Opensource Library.
Keywords : Reinforcement Learning, FPS Game, ViZDoom, Unity ML Agents, A2C, Q-Learning.
