Vicki Allan 2008. Looking for students for two NSF funded grants. Funded Projects 2008-2011. CPATH – Computing Concepts Educational Curriculum Development Looking for help in the creation of a new introductory course – USU 1360 COAL – Coalition Formation Research in Multi-agent systems.
Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.
Looking for students
for two NSF funded grants
The balls on the left are to be exchanged Modules (ILMs)
with the balls on the right by a sequence
of moves. Any ball can move into adjacent empty slot. Any ball can jump over a single neighbor to an empty slot.
Abstraction – general purpose rules
Second project involves multi-agent systems
Binary Protocol Modules (ILMs)
One voter ranks c > d > b > a
One voter ranks a > c > d > b
One voter ranks b > a > c > d
winner (c, (winner (a, winner(b,d)))=a
winner (d, (winner (b, winner(c,a)))=d
winner (d, (winner (c, winner(a,b)))=c
winner (b, (winner (d, winner(c,a)))=b
surprisingly, order of pairing yields different winner!
a=19, b=24, c=17, d=10
Just counting first ranks isn’t enough.
Borda protocol count the number of times a person was ranked first?
assigns an alternative |O| points for the highest preference, |O|-1 points for the second, and so on
reasonable??? count the number of times a person was ranked first?
a=18, b=19, c=20, d=13
Is this a good way?
a=18, b=19, c=20, d=13
When loser is removed, second worst becomes winner!
Vicki Allan – Utah State University count the number of times a person was ranked first?
Kevin Westwood – Utah State University
Presented September 2007, Netherlands
(Work also presented in Hong Kong, Finland, Australia, California)
Who Works Together in Agent Coalition Formation?
Inspired by RFP (Kraus, Shehory, and Taase 2003)
determined probabilistically – no two agents alike.
1) Randomly orders Agents
2) Each agent gets a turn
3) Coalitions are awarded task
Only Change in coalition
How do I decide what to propose? count the number of times a person was ranked first?
If I make an offer…
If others have made me an offer…
1) Task Selection
2) Other Agents Selected for Team
Competitive: best for me
Why not always be greedy?
Wouldn’t this always be a noble thing to do?
Perhaps no one else can do it
Maybe it shouldn’t be done
Melting – same deal gone later
Some amount of compromise is necessary… count the number of times a person was ranked first?
We term the fraction of the total possible you demand – the compromising ratio
Task Rich: 2 tasks count the number of times a person was ranked first?
for every agent
My tasks parallel total tasks
Dilemma for second tier university
Note how profit is affect by load count the number of times a person was ranked first?
Only Local Profit agents
change compromising ratio
Yet others slightly increase too
For every agent type, the most likely proposer count the number of times a person was ranked first?
was a Local Profit agent.
but Local Profit agent doesn’t accept
Coopetitive proposals especially well
For every agent type, proposals,
Best Fit is a strong acceptor.
Perhaps because it isn’t accepted well as a proposer
Load balance seems to affect roles proposals,
Coopetitive agents function better as proposers to Local Profit agents in balanced or task rich environment.
Coopetitive Agents look
at fit as long as it isn’t too bad
compared to profit.
Coopetitive accepts most proposals
from agents like itself
in agent rich environments
Best fit does better and better as more dominant in set
shows relationship if all equal percent
does better when
it isn’t dominant