Learning to Coordinate

with Coordination Graphs

in Repeated

Single-Stage Multi-Agent

Decision Problems

Eugenio Bargiacchi, Timothy Verstraeten, Diederik Roijers, Ann Nowé, Hado van Hasselt

Learning to Coordinate with Coordination Graphs in Repeated Single-Stage Multi-Agent Decision Problems Eugenio Bargiacchi, Timothy Verstraeten, Diederik Roijers, Ann Nowé, Hado van Hasselt

Learning to Coordinate

with Coordination Graphs

in Repeated

Single-Stage Multi-Agent

Decision Problems

Possible approach: UCB

Does not scale with AN Exponential NUMBER OF joint actions

+

OUR CONTRIBUTION: MAUCE

MAXIMIZE:

OUR CONTRIBUTION: MAUCE

WE INTRODUCE UCVE, A VARIABLE ELIMINATION-type Algorithm

prune suboptimal local joint actions, reducing the number of joint actions to consider.

variable elimination

variable elimination

variable elimination

variable elimination

variable elimination

OUR CONTRIBUTION: MAUCE

WE INTRODUCE UCVE, A VARIABLE ELIMINATION-type Algorithm

prune suboptimal local joint actions, reducing the number of joint actions to consider.

OUR CONTRIBUTION: MAUCE

Exploration bonus bounds regret

linear in the number of agents!

Experiments

All code released open source!

QUestions?

thank you!

BENERL 2018

BENERL 2018

svalorzen

Learning to Coordinate

with Coordination Graphs

in Repeated

Single-Stage Multi-Agent

Decision Problems

BENERL 2018

More from svalorzen