Pooja Bharat Raskar
Rahi Vilas Thale
Krutika Satish Shinde
Nikita Saindane
- Student Department of Computer Engineering, Pillai HOC College of Engineering and Technology, Rasayani Maharashtra India
- Student Department of Computer Engineering, Pillai HOC College of Engineering and Technology, Rasayani Maharashtra India
- Student Department of Computer Engineering, Pillai HOC College of Engineering and Technology, Rasayani Maharashtra India
- Assistant Professor Department of Computer Engineering, Pillai HOC College of Engineering and Technology, Rasayani Maharashtra India
Abstract
It is essential for research on computational and/or artificial intelligence (CI/AI) applied to games to have relevant games to apply AI algorithms to. This is pertinent. It doesn’t matter if one is studying how to use CI/AI techniques to test and improve AI (e.g., games provide challenging yet scalable problems which engage many central aspects of human cognitive capacity) or how to use CI/AI techniques to improve games (e.g., player satisfaction modeling, procedural content generation, and the creation of believable and interesting bots).. Using reinforcement learning techniques, this paper proposes a thorough framework for teaching an intelligent agent to play Mario, a beloved video game. To improve game performance, the suggested model integrates the Proximal Policy Optimization (PPO) algorithm with the prospective integration of the Deep Q-Learning (DQL) algorithm. Setting up Mario in a gaming environment, starting the game, and preparing the game state are all part of the setup process. By utilizing vectorization and grayscale approaches, the agent is able to depict the game environment more effectively. To improve our method, we take cues from seminal academic articles like “A Survey of Deep Reinforcement Learning in Video Games” and “System Design for an Integrated Lifelong Reinforcement Learning Agent for Real-Time Strategy Games.”
The research looks at PPO and DQL’s individual effectiveness as well as comparing their performance to reveal their advantages and disadvantages when it comes to teaching an AI agent to overcome the obstacles in the Mario game.
Keywords: AI Model, Game Reinforcement Learning, Proximal Policy Optimization, Artificial intelligence, Mario Game
References
[1] Variations on the Reinforcement Learning performance of Blackjack, Avish Buramdoyal, Tim Gebbie, arXiv, 9 Aug 2023
[2] System Design for an Integrated Lifelong Reinforcement Learning Agent for Real-Time Strategy Games, Indranil Sur,
Zachary Daniels, Abrar Rahman, arXiv, 8 Dec 2022
[3] Towards Playing Full MOBA Games with Deep Reinforcement Learning, Deheng Ye, Guibin Chen, Wen Zhang, Sheng Chen, arXiv, 31 Dec 2020
[4] A Survey of Deep Reinforcement Learning in Video Games,Kun Shao, Zhentao Tang, Yuanheng Zhu, Member, IEEE, Nannan Li, and Dongbin Zhao, Fellow, IEEE, 26 Dec 2019.
[5] Sah SK, Fidele H. Enhancing Mario Gaming Using Optimized Reinforcement Learning. InInternational Conference on Distributed Computing and Intelligent Technology 2024 Jan 4 (pp. 206-221). Cham: Springer Nature Switzerland.
[6] Karakovskiy S, Togelius J. The mario ai benchmark and competitions. IEEE Transactions on Computational Intelligence and AI in Games. 2012 Feb 22;4(1):55-67.
[7] Lee G, Luo M, Zambetta F, Li X. Learning a Super Mario controller from examples of human play. In2014 IEEE Congress on Evolutionary Computation (CEC) 2014 Jul 6 (pp. 1-8). IEEE.
[8] Mohan S, Laird JE. Learning to play Mario. Tech. Rep. CCA-TR-2009-03. 2009.
[9] Engelsvoll RN, Gammelsrød A, Thoresen BI. Generating levels and playing Super Mario Bros. with deep reinforcement learning using various techniques for level generation and deep q-networks for playing (Master’s thesis, University of Agder).
[10] Hou NC, Hong NS, On CK, Teo J. Infinite Mario Bross AI using genetic algorithm. In2011 IEEE Conference on Sustainable Utilization and Development in Engineering and Technology (STUDENT) 2011 Oct 20 (pp. 85-89). IEEE.
[11] Arzate Cruz C, Igarashi T. Mariomix: Creating aligned playstyles for bots with interactive reinforcement learning. InExtended abstracts of the 2020 annual symposium on computer-human interaction in play 2020 Nov 2 (pp. 134-139).
[12] Buramdoyal A, Gebbie T. Variations on the Reinforcement Learning performance of Blackjack. arXiv preprint arXiv:2308.07329. 2023 Aug 9.
[13] Sur I, Daniels Z, Rahman A, Faber K, Gallardo G, Hayes T, Taylor C, Gurbuz MB, Smith J, Joshi S, Japkowicz N. System design for an integrated lifelong reinforcement learning agent for real-time strategy games. InProceedings of the Second International Conference on AI-ML Systems 2022 Oct 12 (pp. 1-9).
[14] Ye D, Chen G, Zhang W, Chen S, Yuan B, Liu B, Chen J, Liu Z, Qiu F, Yu H, Yin Y. Towards playing full moba games with deep reinforcement learning. Advances in Neural Information Processing Systems. 2020;33:621-32.
[15] Shao K, Tang Z, Zhu Y, Li N, Zhao D. A survey of deep reinforcement learning in video games. arXiv preprint arXiv:1912.10944. 2019 Dec 23.
Journal of Instrumentation Technology & Innovations
| Volume | |
| Received | June 14, 2024 |
| Accepted | June 28, 2024 |
| Published | July 16, 2024 |
function myFunction2() {
var x = document.getElementById(“browsefigure”);
if (x.style.display === “block”) {
x.style.display = “none”;
}
else { x.style.display = “Block”; }
}
document.querySelector(“.prevBtn”).addEventListener(“click”, () => {
changeSlides(-1);
});
document.querySelector(“.nextBtn”).addEventListener(“click”, () => {
changeSlides(1);
});
var slideIndex = 1;
showSlides(slideIndex);
function changeSlides(n) {
showSlides((slideIndex += n));
}
function currentSlide(n) {
showSlides((slideIndex = n));
}
function showSlides(n) {
var i;
var slides = document.getElementsByClassName(“Slide”);
var dots = document.getElementsByClassName(“Navdot”);
if (n > slides.length) { slideIndex = 1; }
if (n (item.style.display = “none”));
Array.from(dots).forEach(
item => (item.className = item.className.replace(” selected”, “”))
);
slides[slideIndex – 1].style.display = “block”;
dots[slideIndex – 1].className += ” selected”;
}