Training a Negotiating Agent through Self-Play

Training a Negotiating Agent through Self-Play

Jurševskis, Renāts (TU Delft Electrical Engineering, Mathematics and Computer Science)

Renting, B.M. (mentor)
Murukannaiah, P.K. (mentor)
Zhang, X. (graduation committee)

Delft University of Technology

Computer Science and Engineering

CSE3000 Research Project

2022-06-23

Recent developments in applying reinforcement learning to cooperative environments, like negotiation, have brought forward an important question: how well can a negotiating agent be trained through self-play? Previous research has seen successful application of self-play to other settings, like the games of chess and Go. This paper explores the usage of self-play within the training of a negotiating agent and determines if it is possible to successfully train an agent purely through self-play. The results of the experimentation show that a training stage using self-play can match or even exceed an approach using a set of training opponents. By using multiple self-play opponents, the average utility can be further improved by introducing more variance during training. In addition, using a combination of both self-play and training opponents leads to a hybrid approach that performs better than either of the two techniques separately.

negotiating agents
self-play
reinforcement learning

http://resolver.tudelft.nl/uuid:4b1cc402-cb64-4c86-ac5b-6fcb891f8472

Student theses

bachelor thesis

© 2022 Renāts Jurševskis