Abstract
We propose a variable reward scheme in decentralized multi-agent deep reinforcement learning for a sequential task consisting of a number of subtasks which can be completed when all subtasks are executed in a certain order before a deadline by agents with different capabilities. Developments in computer science and robotics are drawing attention to multi-agent systems for complex tasks. However, coordinated behavior among agents requires sophistication and is highly dependent on the structures of tasks and environments; thus, it is preferable to individually learn appropriate coordination depending on specific tasks. This study focuses on the learning of a sequential task by cooperative agents from a practical perspective. In such tasks, agents must learn both efficiency for their own subtasks and coordinated behavior for other agents because the former provides more chances for the subsequent agents to learn, while the latter facilitates the execution of subsequent subtasks. Our proposed reward scheme enables agents to learn these behaviors in a balanced manner. We then experimentally show that agents in the proposed reward scheme can achieve more efficient task execution compared to baseline methods based on static reward schemes. We also analyzed the learned coordinated behavior to see the reasons of efficiency.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Chang, Y.H., Ho, T., Kaelbling, L.P.: All learning is local: multi-agent learning in global reward games. In: Proceedings of the 16th International Conference on Neural Information Processing Systems, pp. 807–814. NIPS2003, MIT Press, Cambridge, MA, USA (2003)
Diallo, E.A.O., Sugiyama, A., Sugawara, T.: Learning to coordinate with deep reinforcement learning in doubles pong game. In: 2017 16th IEEE International Conference on Machine Learning and Applications (ICMLA), pp. 14–19. IEEE (2017)
Du, Y., Han, L., Fang, M., Liu, J., Dai, T., Tao, D.: LIIR: Learning individual intrinsic reward in multi-agent reinforcement learning. In: Wallach, H., Larochelle, H., Beygelzimer, A., d’Alché-Buc, F., Fox, E., Garnett, R. (eds.) Advances in Neural Information Processing Systems, vol. 32. Curran Associates Inc. (2019)
Foerster, J.N., Assael, Y.M., de Freitas, N., Whiteson, S.: Learning to communicate with deep multi-agent reinforcement learning. In: Proceedings of the 30th International Conference on Neural Information Processing Systems, pp. 2145–2153. NIPS2016, Curran Associates Inc., Red Hook, NY, USA (2016)
Lowe, R., Wu, Y., Tamar, A., Harb, J., Abbeel, P., Mordatch, I.: Multi-agent actor-critic for mixed cooperative-competitive environments. In: Proceedings of the 31st International Conference on Neural Information Processing Systems, pp. 6382–6393. NIPS2017, Curran Associates Inc., Red Hook, NY, USA (2017)
Miyashita, Y., Sugawara, T.: Coordinated behavior for sequential cooperative task using two-stage reward assignment with decay. In: Yang, H., Pasupa, K., Leung, A.C.-S., Kwok, J.T., Chan, J.H., King, I. (eds.) ICONIP 2020. LNCS, vol. 12533, pp. 257–269. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-63833-7_22
Mnih, V., et al.: Playing atari with deep reinforcement learning. arXiv preprint arXiv:1312.5602 (2013)
Ng, A.Y., Harada, D., Russell, S.J.: Policy invariance under reward transformations: theory and application to reward shaping. In: Proceedings of the 16th International Conference on Machine Learning, pp. 278–287 (1999)
Ng, A.Y., Russell, S.J.: Algorithms for inverse reinforcement learning. In: Proceedings of the 17th International Conference on Machine Learning, pp. 663–670 (2000)
Tampuu, A., et al.: Multiagent cooperation and competition with deep reinforcement learning. PLoS ONE 12(4), e0172395 (2017)
Tan, M.: Multi-agent reinforcement learning: independent vs. cooperative agents. In: Proceedings of the Tenth International Conference on Machine Learning, pp. 330–337 (1993)
Van Hasselt, H., Guez, A., Silver, D.: Deep reinforcement learning with double Q-learning. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 30 (2016)
Vinyals, O., et al.: Grandmaster level in StarCraft II using multi-agent reinforcement learning. Nature 575(7782), 350–354 (2019)
Zhou, M., Liu, Z., Sui, P., Li, Y., Chung, Y.Y.: Learning implicit credit assignment for cooperative multi-agent reinforcement learning. In: Advance in Neural Information Processing Systems, vol. 33, pp. 11853–11864. Curran Associates Inc., (2020)
Author information
Authors and Affiliations
Corresponding authors
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Oguni, Y., Miyashita, Y., Sugawara, T. (2022). Shifting Reward Assignment for Learning Coordinated Behavior in Time-Limited Ordered Tasks. In: Dignum, F., Mathieu, P., Corchado, J.M., De La Prieta, F. (eds) Advances in Practical Applications of Agents, Multi-Agent Systems, and Complex Systems Simulation. The PAAMS Collection. PAAMS 2022. Lecture Notes in Computer Science(), vol 13616. Springer, Cham. https://doi.org/10.1007/978-3-031-18192-4_24
Download citation
DOI: https://doi.org/10.1007/978-3-031-18192-4_24
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-18191-7
Online ISBN: 978-3-031-18192-4
eBook Packages: Computer ScienceComputer Science (R0)