RL4CO: An Extensive Reinforcement Learning for Combinatorial Optimization Benchmark
Publikation: Beiträge in Sammelwerken › Aufsätze in Konferenzbänden › Forschung › begutachtet
Standard
KDD 2025 - Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining. Hrsg. / Luiza Antonie; Jian Pei; Xiaohui Yu. Association for Computing Machinery, 2025. S. 5278-5289 (Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining; Band 2).
Publikation: Beiträge in Sammelwerken › Aufsätze in Konferenzbänden › Forschung › begutachtet
Harvard
APA
Vancouver
Bibtex
}
RIS
TY - CHAP
T1 - RL4CO
T2 - 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD 2025
AU - Berto, Federico
AU - Hua, Chuanbo
AU - Park, Junyoung
AU - Luttmann, Laurin
AU - Ma, Yining
AU - Bu, Fanchen
AU - Wang, Jiarui
AU - Ye, Haoran
AU - Kim, Minsu
AU - Choi, Sanghyeok
AU - Zepeda, Nayeli Gast
AU - Hottung, André
AU - Zhou, Jianan
AU - Bi, Jieyi
AU - Hu, Yu
AU - Liu, Fei
AU - Kim, Hyeonah
AU - Son, Jiwoo
AU - Kim, Haeyeon
AU - Angioni, Davide
AU - Kool, Wouter
AU - Cao, Zhiguang
AU - Zhang, Qingfu
AU - Kim, Joungho
AU - Zhang, Jie
AU - Shin, Kijung
AU - Wu, Cathy
AU - Ahn, Sungsoo
AU - Song, Guojie
AU - Kwon, Changhyun
AU - Tierney, Kevin
AU - Xie, Lin
AU - Park, Jinkyoo
N1 - Publisher Copyright: © 2025 Association for Computing Machinery. All rights reserved.
PY - 2025/8/3
Y1 - 2025/8/3
N2 - Combinatorial optimization (CO) is fundamental to several real-world applications, from logistics and scheduling to hardware design and resource allocation. Deep reinforcement learning (RL) has recently shown significant benefits in solving CO problems, reducing reliance on domain expertise and improving computational efficiency. However, the absence of a unified benchmarking framework leads to inconsistent evaluations, limits reproducibility, and increases engineering overhead, raising barriers to adoption for new researchers. To address these challenges, we introduce RL4CO, a unified and extensive benchmark with in-depth library coverage of 27 CO problem environments and 23 state-of-the-art baselines. Built on efficient software libraries and best practices in implementation, RL4CO features modularized implementation and flexible configurations of diverse environments, policy architectures, RL algorithms, and utilities with extensive documentation. RL4CO helps researchers build on existing successes while exploring and developing their own designs, facilitating the entire research process by decoupling science from heavy engineering. We finally provide extensive benchmark studies to inspire new insights and future work. RL4CO has already attracted numerous researchers in the community and is open-sourced at https://github.com/ai4co/rl4co.
AB - Combinatorial optimization (CO) is fundamental to several real-world applications, from logistics and scheduling to hardware design and resource allocation. Deep reinforcement learning (RL) has recently shown significant benefits in solving CO problems, reducing reliance on domain expertise and improving computational efficiency. However, the absence of a unified benchmarking framework leads to inconsistent evaluations, limits reproducibility, and increases engineering overhead, raising barriers to adoption for new researchers. To address these challenges, we introduce RL4CO, a unified and extensive benchmark with in-depth library coverage of 27 CO problem environments and 23 state-of-the-art baselines. Built on efficient software libraries and best practices in implementation, RL4CO features modularized implementation and flexible configurations of diverse environments, policy architectures, RL algorithms, and utilities with extensive documentation. RL4CO helps researchers build on existing successes while exploring and developing their own designs, facilitating the entire research process by decoupling science from heavy engineering. We finally provide extensive benchmark studies to inspire new insights and future work. RL4CO has already attracted numerous researchers in the community and is open-sourced at https://github.com/ai4co/rl4co.
KW - benchmark
KW - combinatorial optimization
KW - neural combinatorial optimization
KW - open research community
KW - reinforcement learning
KW - Business informatics
UR - http://www.scopus.com/inward/record.url?scp=105014477776&partnerID=8YFLogxK
U2 - 10.1145/3711896.3737433
DO - 10.1145/3711896.3737433
M3 - Article in conference proceedings
AN - SCOPUS:105014477776
T3 - Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining
SP - 5278
EP - 5289
BT - KDD 2025 - Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining
A2 - Antonie, Luiza
A2 - Pei, Jian
A2 - Yu, Xiaohui
PB - Association for Computing Machinery
Y2 - 3 August 2025 through 7 August 2025
ER -