Using Dialogue Games to Form Coalitions with Self-Interested Agents Luke Riley Department of Computer Science University of Liverpool

Slides:



Advertisements
Similar presentations
TWO STEP EQUATIONS 1. SOLVE FOR X 2. DO THE ADDITION STEP FIRST
Advertisements

1 Copyright © 2010, Elsevier Inc. All rights Reserved Fig 2.1 Chapter 2.
By D. Fisher Geometric Transformations. Reflection, Rotation, or Translation 1.
1 Probability and the Web Ken Baclawski Northeastern University VIStology, Inc.
Source of slides: Introduction to Automata Theory, Languages and Computation.
Business Transaction Management Software for Application Coordination 1 Business Processes and Coordination.
Energy-Efficient Distributed Algorithms for Ad hoc Wireless Networks Gopal Pandurangan Department of Computer Science Purdue University.
Jeopardy Q 1 Q 6 Q 11 Q 16 Q 21 Q 2 Q 7 Q 12 Q 17 Q 22 Q 3 Q 8 Q 13
Jeopardy Q 1 Q 6 Q 11 Q 16 Q 21 Q 2 Q 7 Q 12 Q 17 Q 22 Q 3 Q 8 Q 13
Title Subtitle.
Coordinate Plane Practice The following presentation provides practice in two skillsThe following presentation provides practice in two skills –Graphing.
0 - 0.
DIVIDING INTEGERS 1. IF THE SIGNS ARE THE SAME THE ANSWER IS POSITIVE 2. IF THE SIGNS ARE DIFFERENT THE ANSWER IS NEGATIVE.
MULTIPLYING MONOMIALS TIMES POLYNOMIALS (DISTRIBUTIVE PROPERTY)
ADDING INTEGERS 1. POS. + POS. = POS. 2. NEG. + NEG. = NEG. 3. POS. + NEG. OR NEG. + POS. SUBTRACT TAKE SIGN OF BIGGER ABSOLUTE VALUE.
MULTIPLICATION EQUATIONS 1. SOLVE FOR X 3. WHAT EVER YOU DO TO ONE SIDE YOU HAVE TO DO TO THE OTHER 2. DIVIDE BY THE NUMBER IN FRONT OF THE VARIABLE.
SUBTRACTING INTEGERS 1. CHANGE THE SUBTRACTION SIGN TO ADDITION
MULT. INTEGERS 1. IF THE SIGNS ARE THE SAME THE ANSWER IS POSITIVE 2. IF THE SIGNS ARE DIFFERENT THE ANSWER IS NEGATIVE.
Addition Facts
1 Student-Project Allocation with Preferences over Projects David Manlove Gregg OMalley University of Glasgow Department of Computing Science Supported.
Module 2 Sessions 10 & 11 Report Writing.
ZMQS ZMQS
Artificial Bee Colony Algorithm
Real-Time Competitive Environments: Truthful Mechanisms for Allocating a Single Processor to Sporadic Tasks Anwar Mohammadi, Nathan Fisher, and Daniel.
Chapter 18 Methodology – Monitoring and Tuning the Operational System Transparencies © Pearson Education Limited 1995, 2005.
Shadow Prices vs. Vickrey Prices in Multipath Routing Parthasarathy Ramanujam, Zongpeng Li and Lisa Higham University of Calgary Presented by Ajay Gopinathan.
The Weighted Proportional Resource Allocation Milan Vojnović Microsoft Research Joint work with Thành Nguyen Microsoft Research Asia, Beijing, April, 2011.
INTRODUCTION TO SIMULATION WITH OMNET++ José Daniel García Sánchez ARCOS Group – University Carlos III of Madrid.
© S Haughton more than 3?
What is Pay & Performance?
Chapter 2 Section 3.
Linking Verb? Action Verb or. Question 1 Define the term: action verb.
© 2007 Lawrenceville Press Slide 1 Chapter 3 Margins.
Squares and Square Root WALK. Solve each problem REVIEW:
This, that, these, those Number your paper from 1-10.
Reaching Agreements II. 2 What utility does a deal give an agent? Given encounter  T 1,T 2  in task domain  T,{1,2},c  We define the utility of a.
1 First EMRAS II Technical Meeting IAEA Headquarters, Vienna, 19–23 January 2009.
Addition 1’s to 20.
25 seconds left…...
Test B, 100 Subtraction Facts
Week 1.
We will resume in: 25 Minutes.
Dantzig-Wolfe Decomposition
1 Unit 1 Kinematics Chapter 1 Day
Immunobiology: The Immune System in Health & Disease Sixth Edition
1 PART 1 ILLUSTRATION OF DOCUMENTS  Brief introduction to the documents contained in the envelope  Detailed clarification of the documents content.
How Cells Obtain Energy from Food
Auction Theory Class 5 – single-parameter implementation and risk aversion 1.
A Short Tutorial on Cooperative Games
Goals Become familiar with current research topics
NON - zero sum games.
Indiana University’s Remote Classroom Click to edit Master text styles Second level Third level Fourth level Fifth level Non-science majors learn about.
Negotiating a stable distribution of the payoff among agents may prove challenging. The issue of coalition formation has been investigated extensively,
Behaviour.
Using Dialogue Games to Form Coalitions with Self-Interested Agents Luke Riley Department of Computer Science University of Liverpool
Complexity of Determining Nonemptiness of the Core Vincent Conitzer, Tuomas Sandholm Computer Science Department Carnegie Mellon University.
Computing Shapley values, manipulating value division schemes, and checking core membership in multi-issue domains Vincent Conitzer, Tuomas Sandholm Computer.
Complexity of Determining Nonemptiness of the Core
Click to edit Master text styles
Author names here Author association names here
Click to edit Master text styles
Click to edit Master text styles
ОПШТЕСТВО ТЕМА: МЕСТОТО ВО КОЕ ЖИВЕАМ Скопје
Author names here Author associations here
Author names here Author associations here
Click to edit Master text styles
Author names here Author associations here
Click to edit Master text styles
Presentation transcript:

Using Dialogue Games to Form Coalitions with Self-Interested Agents Luke Riley Department of Computer Science University of Liverpool Supervisors: Katie Atkinson & Terry Payne

Click to edit the outline text format Second Outline Level Third Outline Level Fourth Outline Level Fifth Outline Level Sixth Outline Level Seventh Outline Level Eighth Outline Level Ninth Outline LevelClick to edit Master text styles Second level Third level Fourth level Fifth level 2 1. Coalition Formation in Cooperative Game Theory. 2. Coalition Formation in Argumentation. 3. The Issues and Problems Between these Two Approaches. 4. My Research.

3 1. Coalition Formation in Cooperative Game Theory (CGT)

4 Background N-person cooperative games (coalition games) were proposed in 1944 by von Neumann & Morgenstern [1]: Where... [1] J. von Neumann and O. Morgenstern. The Theory of Games and Economic Behavior. Princeton University Press, Characteristic Function: Agent set:

5 In its most traditional style the CGT outcome of a coalition game is: Where... CS = a set of coalitions (the coalition structure) x = a vector of each individual agent's payoff in the game. Solving a Coalition Game

6 Finding a Stable Outcome – The Core A Coalition Structure is core-stable if no subset of agents can benefit from defecting to another coalition. The core [2] is the set where: e.g. Example 1 : Given a coalition game where N = {1,2}, v({1}) = v({2}) = 5 and v({1,2}) = 20 the proposed core outcome is e.g. Example 2 : Given a coalition game where N = {1,2}, v({1}) = v({2}) = 5 and v({1,2}) = 20 the proposed core outcome is Yet core payoffs can sometimes be unfair [2] D. Gillies. Some theorems on n-person games. PhD thesis, Princeton University, 1953.

7 Epsilon-Core Solution [3] The epsilon value can be seen as the cost of deviating. Also the core can sometimes be empty e.g. Example 3 : Given a coalition game where N = {1,2,3}, forall subsets C if |C| = 2 then v(C) = 1 else v(C) = 0 e.g. Example 4 : Given the coalition game of example 3, the payoff vector x(1/3,1/3,1/3) is 1/3- core stable. [3] Shapley, Lloyd S. and Shubik, M. Quasi-cores in a monetary economy with non-convex preferences, Econometrica (The Econometric Society) 34(4): 805–827, 1966.

8 1. Coalition Formation in Argumentation

9 Dung's Initial Work Dung showed that Argumentation Frameworks were natural ways to represent n-person games, for example theorem 6 of [4]: [4] P. M. Dung. On the acceptability of arguments and its fundamental role in nonmonotonic reasoning, logic programming and n-person games. Artificial Intelligence, 77:321–357, x(3,4,8) x(3,3,5)x(3,3,3) The AF represents 3 possible payoff vectors of the coalition game: v({1}) = v({2}) = v({3}) = 3 v({1,3}) = 8 v({2,3}) = 12 or v(C) = 0

10 Amgoud's Further Research Amgoud in [5] extended this research, where she highlighted: How to always find a solution to a coalition game Outlines how agents can collaboratively build AFs for coalition games How a dialogue game can be used to check if a certain coalition was in the best coalition structure [5] L. Amgoud. An argumentation-based model for reasoning about coalition structures. In ArgMAS, pages , 2005.

11 2. The Issues of Joining the Two Approaches

12 Various Issues CGT: Lacks flexible communication protocols to form stable coalitions. CGT: Generally does not take into account the computation and communication costs of finding stable coalition structures from a MAS perspective. Arg: There is little research showing how payoff vectors are found and justified by MAS. Arg: No research on how to stabilise coalitions games, using the epsilon-core Arg: Only some limited direct mapping between the argumentation models and the CGT coalition game types (e.g. static, dynamic, skill games,...)

13 My Current Research Question How can self-interested agents make use of argumentation within their communication to enable them to form a stable optimal coalition structure with an approximately fair payoff distribution?

14 3. The Proposed Method

15 Dialogue Games & Argumentation Schemes Dialogue Games can be used to build argumentation frameworks in real time, where agents can assert and retract arguments. Argumentation schemes are patterns of reasoning that when instantiated provide presumptive justification for the particular conclusion of the scheme e.g:...

16 Approximately fair payoffs AFs can easily represent the core...But the core can be unfair Solution – restrict the payoffs allowed: agents have to propose an equal split of v(C) or each agent should be given at least the same value it can get from a coalition of agents willing to defect Agents can object to a proposed payoff by finding a better one for itself. Once a core payoff is found, the dialogue stops

17 Dialogue Games & Argumentation Schemes I have devised a dialogue game [6] to find an optimal coalition structure with a restricted core payoff Moves: e.g: [6] L. Riley, K. Atkinson, and T. Payne. Coalition structure generation for self interested agents in a dialogue game. Technical Report ULCS , University of Liverpool, 2012.

18 Core example MoveCoalition{1}{2}{3}{1,2}{1,3}{2,3}{1,2,3} Coalition value [3][9/9] 2[2][10/4] 3[3][11/7] FINISH Coalition Structure of move 3 is {{1,3}, {2}}, the payoff vector is x(11,3,7) and is core stable

19 Epsilon-Core Example Movee valueCoalition{1}{2}{3}{1,2}{1,3}{2,3}{1,2,3} value [5][11/11] 1value [6][7/12] 2value [7][8/8] 3value [8][9.5/9.5] 4value FINISH Coalition Structure of move 8 is {{1},{2,3}}, the payoff vector is x(8,9.5,9.5) and is 3-core stable

20 Potential Future Modify argumentation scheme and attack relations so that other coalition games can be modeled (e.g stochastic, dynamic, skill games,...). Optimise process: Combine mechanism design approach of [7] with efficient distribution methods of [8]. [7] Tuomas Sandholm, Kate Larson, Martin Andersson, Onn Shehory and Fernando Tohmé, Coalition structure generation with worst case guarantees, Artificial Intelligence, Volume 111, Issues 1–2, July 1999, Pages [8] T. Rahwan. Algorithms for Coalition Formation in Multi-Agent Systems. PhD thesis, University of Southampton, 2007.

21 Thanks For Listening Questions?