• Version
• Download 23
• File Size 293.80 KB
• File Count 1
• Create Date 10 March 2026
• Last Updated 10 March 2026

Game Mind: Reinforcement Learning Framework for Autonomous Agents

1K. Parameshwara Rao, 2N. Sreeja, 3S. Srikanth, 4Sr. Asst. Prof.CH. Sushma

123Bachelors in Computer Science and Engineering AIML, Geethanjali College of Engineering and Technology, Hyderabad, Telangana, India

4Senior Assistant Professor, Computer Science, Geethanjali College of Engineering and Technology, Hyderabad, Telangana, India

ABSTRACT: In today's world, big tech companies like IBM and Boston invest heavily in training their robots in various types of real- time environments, a process that is both costly and time-consuming. Despite these efforts, agents won't be fully prepared for real-world scenarios due to the limited training environments. To overcome these constraints, it is essential to expose AI agents to a broader range of simulated conditions. Thus, Virtual training of AI agents is the best way to increase their efficiency. In this project, an AI-Driven Game Bot that uses Reinforcement Learning (RL) techniques to learn and compete on its own in simulated game environments is developed. Through constant interaction with its surroundings and rewards for reaching predetermined objectives, the agent learns optimal behaviors rather than depending on preset strategies. Multiple AI agents can train cooperatively or competitively in dynamic 3D simulations like soccer, racing, or battle arenas thanks to the model's extensibility beyond a single game.

Keywords: Reinforcement Learning, Game AI, Autonomous Agents, 3D Simulation, Multi- Agent System, Intelligent Bot, Self-Learning, Environment Interaction, Adaptive Gameplay