Kagan Tumer's Publications

Display Publications by [Year] [Type] [Topic]


Multi-Objective Multiagent Credit Assignment Through Difference Rewards in Reinforcement Learning. L. Yliniemi and K. Tumer. In The Tenth International Conference on Simulated Evolution And Learning (SEAL 2014), Dunedin, New Zealand, December 2014.

Abstract

Multiagent systems have had a powerful impact on the real world. Many of the systems it studies (air traffic, satellite coordination, rover exploration) are inherently multi-objective, but they are often treated as single-objective problems within the research. A very important concept within multiagent systems is that of credit assignment: clearly quantifying an individual agentŐs impact on the overall system performance. In this work we extend the concept of credit assignment into multi-objective problems, broadening the traditional multiagent learning framework to account for multiple objectives. We show in two domains that by leveraging established credit assignment principles in a multi-objective setting, we can improve performance by (i) increasing learning speed by up to 10x; (ii) reducing sensitivity to unmodeled disturbances by up to 98.4%; and (iii) producing solutions that dominate all solutions discovered by a traditional team- based credit assignment schema. Our results suggest that in a multiagent multi- objective problem, proper credit assignment is as important to performance as the choice of multi-objective algorithm.

Download

[PDF]2.1MB  

BibTeX Entry

@inproceedings{tumer-yliniemi_moma-seal14,
        author = {L. Yliniemi  and  K. Tumer},
        title = {Multi-Objective Multiagent Credit Assignment Through Difference Rewards in Reinforcement Learning},
        booktitle = {The Tenth International Conference on Simulated Evolution And Learning (SEAL 2014)},
	month = {December},
	address = {Dunedin,  New Zealand},
	abstract = {Multiagent systems have had a powerful impact on the real world. Many of the systems it studies (air traffic, satellite coordination, rover exploration) are inherently multi-objective, but they are often treated as single-objective problems within the research. A very important concept within multiagent systems is that of credit assignment: clearly quantifying an individual agentÕs impact on the overall system performance. In this work we extend the concept of credit assignment into multi-objective problems, broadening the traditional multiagent learning framework to account for multiple objectives. We show in two domains that by leveraging established credit assignment principles in a multi-objective setting, we can improve performance by (i) increasing learning speed by up to 10x; (ii) reducing sensitivity to unmodeled disturbances by up to 98.4%; and (iii) producing solutions that dominate all solutions discovered by a traditional team- based credit assignment schema. Our results suggest that in a multiagent multi- objective problem, proper credit assignment is as important to performance as the choice of multi-objective algorithm.},
	bib2html_pubtype = {Refereed Conference Papers},
	bib2html_rescat = {Multiagent Systems, Optimization},
        year = {2014}
        }         

Generated by bib2html.pl (written by Patrick Riley ) on Tue Jun 26, 2018 19:10:42