Comparative Analysis of Reinforcement Learning Algorithms for Autonomous Driving in Simulated 2D Environments: Optimizing Reward Functions and Hyperparameters

Alexander Brunner; Gabriele Kotsis; Ismail Khalil

doi:10.1007/978-3-032-08603-7_8

Comparative Analysis of Reinforcement Learning Algorithms for Autonomous Driving in Simulated 2D Environments: Optimizing Reward Functions and Hyperparameters

Alexander Brunner
, Gabriele Kotsis
, Ismail Khalil^*

^*Corresponding author for this work

Institute of Telecooperation

Research output: Chapter in Book/Report/Conference proceeding › Conference proceedings › peer-review

Abstract

This paper presents a comparative analysis of reinforcement learning (RL) algorithms, specifically Proximal Policy Optimization (PPO) and Deep Q-Network (DQN), for autonomous driving in simulated 2D environments. The study focuses on optimizing reward functions and hyperparameters to enhance road navigation and obstacle avoidance. Our experiments show that DQN generally outperforms PPO in simple environments, while fine-tuning reward structures and hyperparameters significantly impacts the learning process. Techniques such as frame stacking and curriculum learning further improve performance.

Original language	English
Title of host publication	Digital Economy. Emerging Technologies and Business Innovation - 10th International Conference on Digital Economy, ICDEc 2025, Proceedings
Editors	Rim Jallouli, Mohamed Anis Bach Tobji, Nessrine Omrani, Ilyes Jenhani
Publisher	Springer, Cham
Pages	110–125
Number of pages	16
ISBN (Electronic)	978-3-032-08603-7
ISBN (Print)	978-3-032-08602-0
DOIs	https://doi.org/10.1007/978-3-032-08603-7_8
Publication status	E-pub ahead of print - 15 Nov 2025

Publication series

Name	Lecture Notes in Business Information Processing
Volume	560 LNBIP
ISSN (Print)	1865-1348
ISSN (Electronic)	1865-1356

Fields of science

102013 Human-computer interaction
102002 Augmented reality
102006 Computer supported cooperative work (CSCW)
102027 Web engineering
202038 Telecommunications
102021 Pervasive computing
102015 Information systems
102025 Distributed systems
102 Computer Sciences

JKU Focus areas

Digital Transformation
Sustainable Development: Responsible Technologies and Management

Access to Document

10.1007/978-3-032-08603-7_8

Cite this

Brunner, A., Kotsis, G., & Khalil, I. (2025). Comparative Analysis of Reinforcement Learning Algorithms for Autonomous Driving in Simulated 2D Environments: Optimizing Reward Functions and Hyperparameters. In R. Jallouli, M. A. Bach Tobji, N. Omrani, & I. Jenhani (Eds.), Digital Economy. Emerging Technologies and Business Innovation - 10th International Conference on Digital Economy, ICDEc 2025, Proceedings (pp. 110–125). (Lecture Notes in Business Information Processing; Vol. 560 LNBIP). Springer, Cham. Advance online publication. https://doi.org/10.1007/978-3-032-08603-7_8

Brunner, Alexander ; Kotsis, Gabriele ; Khalil, Ismail. / Comparative Analysis of Reinforcement Learning Algorithms for Autonomous Driving in Simulated 2D Environments: Optimizing Reward Functions and Hyperparameters. Digital Economy. Emerging Technologies and Business Innovation - 10th International Conference on Digital Economy, ICDEc 2025, Proceedings. editor / Rim Jallouli ; Mohamed Anis Bach Tobji ; Nessrine Omrani ; Ilyes Jenhani. Springer, Cham, 2025. pp. 110–125 (Lecture Notes in Business Information Processing).

@inproceedings{9ee3703cca744c5d8e79248584948fe9,

title = "Comparative Analysis of Reinforcement Learning Algorithms for Autonomous Driving in Simulated 2D Environments: Optimizing Reward Functions and Hyperparameters",

abstract = "This paper presents a comparative analysis of reinforcement learning (RL) algorithms, specifically Proximal Policy Optimization (PPO) and Deep Q-Network (DQN), for autonomous driving in simulated 2D environments. The study focuses on optimizing reward functions and hyperparameters to enhance road navigation and obstacle avoidance. Our experiments show that DQN generally outperforms PPO in simple environments, while fine-tuning reward structures and hyperparameters significantly impacts the learning process. Techniques such as frame stacking and curriculum learning further improve performance.",

author = "Alexander Brunner and Gabriele Kotsis and Ismail Khalil",

year = "2025",

month = nov,

day = "15",

doi = "10.1007/978-3-032-08603-7\_8",

language = "English",

isbn = "978-3-032-08602-0",

series = "Lecture Notes in Business Information Processing",

publisher = "Springer, Cham",

pages = "110–125",

editor = "Rim Jallouli and \{Bach Tobji\}, \{Mohamed Anis\} and Nessrine Omrani and Ilyes Jenhani",

booktitle = "Digital Economy. Emerging Technologies and Business Innovation - 10th International Conference on Digital Economy, ICDEc 2025, Proceedings",

}

Brunner, A, Kotsis, G & Khalil, I 2025, Comparative Analysis of Reinforcement Learning Algorithms for Autonomous Driving in Simulated 2D Environments: Optimizing Reward Functions and Hyperparameters. in R Jallouli, MA Bach Tobji, N Omrani & I Jenhani (eds), Digital Economy. Emerging Technologies and Business Innovation - 10th International Conference on Digital Economy, ICDEc 2025, Proceedings. Lecture Notes in Business Information Processing, vol. 560 LNBIP, Springer, Cham, pp. 110–125. https://doi.org/10.1007/978-3-032-08603-7_8

Comparative Analysis of Reinforcement Learning Algorithms for Autonomous Driving in Simulated 2D Environments: Optimizing Reward Functions and Hyperparameters. / Brunner, Alexander; Kotsis, Gabriele ; Khalil, Ismail.
Digital Economy. Emerging Technologies and Business Innovation - 10th International Conference on Digital Economy, ICDEc 2025, Proceedings. ed. / Rim Jallouli; Mohamed Anis Bach Tobji; Nessrine Omrani; Ilyes Jenhani. Springer, Cham, 2025. p. 110–125 (Lecture Notes in Business Information Processing; Vol. 560 LNBIP).

Research output: Chapter in Book/Report/Conference proceeding › Conference proceedings › peer-review

TY - GEN

T1 - Comparative Analysis of Reinforcement Learning Algorithms for Autonomous Driving in Simulated 2D Environments: Optimizing Reward Functions and Hyperparameters

AU - Brunner, Alexander

AU - Kotsis, Gabriele

AU - Khalil, Ismail

PY - 2025/11/15

Y1 - 2025/11/15

N2 - This paper presents a comparative analysis of reinforcement learning (RL) algorithms, specifically Proximal Policy Optimization (PPO) and Deep Q-Network (DQN), for autonomous driving in simulated 2D environments. The study focuses on optimizing reward functions and hyperparameters to enhance road navigation and obstacle avoidance. Our experiments show that DQN generally outperforms PPO in simple environments, while fine-tuning reward structures and hyperparameters significantly impacts the learning process. Techniques such as frame stacking and curriculum learning further improve performance.

AB - This paper presents a comparative analysis of reinforcement learning (RL) algorithms, specifically Proximal Policy Optimization (PPO) and Deep Q-Network (DQN), for autonomous driving in simulated 2D environments. The study focuses on optimizing reward functions and hyperparameters to enhance road navigation and obstacle avoidance. Our experiments show that DQN generally outperforms PPO in simple environments, while fine-tuning reward structures and hyperparameters significantly impacts the learning process. Techniques such as frame stacking and curriculum learning further improve performance.

UR - https://www.scopus.com/pages/publications/105022880330

U2 - 10.1007/978-3-032-08603-7_8

DO - 10.1007/978-3-032-08603-7_8

M3 - Conference proceedings

SN - 978-3-032-08602-0

T3 - Lecture Notes in Business Information Processing

SP - 110

EP - 125

BT - Digital Economy. Emerging Technologies and Business Innovation - 10th International Conference on Digital Economy, ICDEc 2025, Proceedings

A2 - Jallouli, Rim

A2 - Bach Tobji, Mohamed Anis

A2 - Omrani, Nessrine

A2 - Jenhani, Ilyes

PB - Springer, Cham

ER -

Brunner A, Kotsis G , Khalil I. Comparative Analysis of Reinforcement Learning Algorithms for Autonomous Driving in Simulated 2D Environments: Optimizing Reward Functions and Hyperparameters. In Jallouli R, Bach Tobji MA, Omrani N, Jenhani I, editors, Digital Economy. Emerging Technologies and Business Innovation - 10th International Conference on Digital Economy, ICDEc 2025, Proceedings. Springer, Cham. 2025. p. 110–125. (Lecture Notes in Business Information Processing). Epub 2025 Nov 15. doi: 10.1007/978-3-032-08603-7_8

Comparative Analysis of Reinforcement Learning Algorithms for Autonomous Driving in Simulated 2D Environments: Optimizing Reward Functions and Hyperparameters

Abstract

Publication series

Fields of science

JKU Focus areas

Access to Document

Other files and links

Cite this