Optimizing DOOM Game Using Reinforcement Learning

doi:10.1007/978-3-032-05545-3_30

2026-01-01

PRE

DOI：10.1007/978-3-032-05545-3_30

原文链接

原文求助

摘要

En 中文

本研究探讨了四种先进强化学习（RL）算法——近端策略优化（PPO）、深度Q网络（DQN）、优势演员-评论家（A2C）和深度确定性策略梯度（DDPG）——在Doom环境中的应用，以优化代理性能，Doom是一款高维第一人称射击游戏。每种算法均被训练以应对不同复杂程度下的导航、目标消除和资源收集等任务。实验整合了定制奖励工程、状态空间设计和环境修改，以挑战代理的适应性。性能评估显示，PPO在连续动作场景中实现了卓越的稳定性和样本效率，而DQN在离散状态-动作空间和确定性奖励方面表现优异。近期应用已证明强化学习在控制游戏代理以实现胜利方面的潜力及其在促进人类开发者游戏开发流程中的作用。A2C通过同时更新演员和评论家网络，展示了强大的可扩展性和快速收敛性。DDPG在学习和连续控制策略方面表现出色，但需要大量的超参数调整以实现最佳性能。在所有实验中，PPO和A2C在处理复杂游戏机制方面表现最一致。对训练时间、计算资源使用和收敛率的比较分析突显了每种方法固有的权衡。研究发现强化学习技术在掌握像Doom这样的动态交互环境中的有效性，并为其实际部署于机器人、自主导航和战略决策系统等现实世界应用提供了见解。

Keyword:

Reinforcement Learning

Proximal Policy Optimization (PPO)

Deep Q-Networks (DQN)

A2C

Deep Deterministic Policy Gradient (DDPG)

Doom Game Optimization

Action Spaces

Agent Training

期刊

PROCEEDINGS OF THE INTERNATIONAL CONFERENCE ON AI AND ROBOTICS, AIR

IF:

论文数:

被引数:

机构

暂无机构信息

引用论文

Intrinsic motivation systems for autonomous mental development 自主心理发展的内在动机系统

IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION

IF12

2007-04-01

687

OAAI

Oudeyer, Pierre-Yves; Kaplan, Frederic; Hafner, Verena V.

Playing FPS Games with Deep Reinforcement Learning

Proceedings of the AAAI Conference on Artificial Intelligence

IF0

2017-02-13

OAAI

Guillaume Lample; Devendra Singh Chaplot

Who Killed Albert Einstein? From Open Data to Murder Mystery Games 谁杀死了阿尔伯特·爱因斯坦？——从开放数据到谋杀谜案游戏

IEEE Transactions on Games

IF0

2019-03-01

OAAI

Gabriella Alves Bulhoes Barros; Michael Cerny Green; Antonios Liapis; Julian Togelius

Human-level control through deep reinforcement learning 通过深度强化学习实现人类层面的控制

Nature

IF0

2015-02-25

PREAI

Volodymyr Mnih; Koray Kavukcuoglu; David Silver; Andrei A. Rusu; Joel Veness; Marc G. Bellemare; Alex Graves; Martin Riedmiller; Andreas K. Fidjeland; Georg Ostrovski; Stig Petersen; Charles Beattie; Amir Sadik; Ioannis Antonoglou; Helen King; Dharshan Kumaran; Daan Wierstra; Shane Legg; Demis Hassabis

Grandmaster level in StarCraft II using multi-agent reinforcement learning 使用多智能体强化学习在星际争霸II中的大师水平

Nature

IF0

2019-10-30

PREAI

Oriol Vinyals; Igor Babuschkin; Wojciech M. Czarnecki; Michaël Mathieu; Andrew Dudzik; Junyoung Chung; David H. Choi; Richard Powell; Timo Ewalds; Petko Georgiev; Junhyuk Oh; Dan Horgan; Manuel Kroiss; Ivo Danihelka; Aja Huang; Laurent Sifre; Trevor Cai; John P. Agapiou; Max Jaderberg; Alexander S. Vezhnevets; Rémi Leblond; Tobias Pohlen; Valentin Dalibard; David Budden; Yury Sulsky; James Molloy; Tom L. Paine; Caglar Gulcehre; Ziyu Wang; Tobias Pfaff; Yuhuai Wu; Roman Ring; Dani Yogatama; Dario Wünsch; Katrina McKinney; Oliver Smith; Tom Schaul; Timothy Lillicrap; Koray Kavukcuoglu; Demis Hassabis; Chris Apps; David Silver

A Survey on Transfer Learning 迁移学习研究综述

IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING

IF10.4

2010-10-01

1.4W

PREAI

Pan, Sinno Jialin; Yang, Qiang

Curiosity-Driven Exploration by Self-Supervised Prediction 通过自我监督的预测进行好奇心驱动的探索

2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)

IF0

2017-07-01

OAAI

Deepak Pathak; Pulkit Agrawal; Alexei A. Efros; Trevor Darrell

Deep learning 深度学习

Nature

IF0

2015-05-27

PREAI

Yann LeCun; Yoshua Bengio; Geoffrey Hinton

Multi-Agent Reinforcement Learning: Independent vs. Cooperative Agents

Machine Learning Proceedings 1993

IF0

1993-01-01

PREAI

Ming Tan

Mastering Atari, Go, chess and shogi by planning with a learned model 通过学习模型进行规划，掌握雅达利，围棋，国际象棋和将棋

Nature

IF0

2020-12-23

PREAI

Julian Schrittwieser; Ioannis Antonoglou; Thomas Hubert; Karen Simonyan; Laurent Sifre; Simon Schmitt; Arthur Guez; Edward Lockhart; Demis Hassabis; Thore Graepel; Timothy Lillicrap; David Silver

没有更多内容