MLKD logo   Machine Learning &
Knowledge Discovery Group

Publication Details

  Author(s): G. Boutsioukis, I. Partalas, I. Vlahavas.

Title: “Transfer Learning in Multi-agent Reinforcement Learning Domains”.

Availability: Click here to download the PDF (Acrobat Reader) file (12 pages).


Appeared in: Accepted for presentation at the 9th European Workshop on Reinforcement Learning and to be published in the Workshop Proceedings, 2011.

Abstract: Transfer learning refers to the process of reusing knowledge from past tasks in order to speed up the learning procedure in new tasks. In reinforcement learning, where agents often require a consider- able amount of training, transfer learning comprises a suitable solution for speeding up learning. Transfer learning methods have primarily been applied in single-agent reinforcement learning algorithms, while no prior work has addressed this issue in the case of multi-agent learning. This work proposes a novel method for transfer learning in multi-agent rein- forcement learning domains. We test the proposed approach in a multi- agent domain under various setups. The results demonstrate that the method helps to reduce the learning time and increase the asymptotic performance.

Relevant Links: