Overlapping Coalition Formation: Charting the Tractability Frontier

34
Overlapping Coalition Formation: Charting the Tractability Frontier Y. Zick, G. Chalkiadakis and E. Elkind (submitted to AAMAS 2012)

description

Overlapping Coalition Formation: Charting the Tractability Frontier. Y. Zick , G. Chalkiadakis and E. Elkind (submitted to AAMAS 2012). Motivation. Agents have limited integer resources. Form Bilateral Trade Contracts : coalitions. The benefit of interaction may be freely divided. - PowerPoint PPT Presentation

Transcript of Overlapping Coalition Formation: Charting the Tractability Frontier

Page 1: Overlapping Coalition Formation: Charting  the Tractability Frontier

Overlapping Coalition Formation: Charting the Tractability Frontier

Y. Zick, G. Chalkiadakis and E. Elkind

(submitted to AAMAS 2012)

Page 2: Overlapping Coalition Formation: Charting  the Tractability Frontier

MotivationAgents have limited

integer resources

The benefit of interaction may be freely divided

Form Bilateral Trade Contracts: coalitions

Page 3: Overlapping Coalition Formation: Charting  the Tractability Frontier

QuestionsWhat is the optimal coalition structure?

How should profits be divided?

Page 4: Overlapping Coalition Formation: Charting  the Tractability Frontier

Problem Complexity

Agents are nodes

The problem can be modeled as a graph

There is an edge between agents if they can profit from collaborating.

Goal: optimal allocation

Page 5: Overlapping Coalition Formation: Charting  the Tractability Frontier

v1(x) = 5I5(x)

v1,2(x,y) = log(x + y + 2)

v2(x) = 0 w1 = 8

w2 = 3

Page 6: Overlapping Coalition Formation: Charting  the Tractability Frontier

v1,2(x,y) = log(x + y + 2)

v2(x) = 0 w1 = 8

w2 = 3

v1(x) = 5I5(x)v1(5) = 5

v1,2(1,1) = 2v1,2(1,1) = 2v1,2(1,1) = 2

Page 7: Overlapping Coalition Formation: Charting  the Tractability Frontier

Computational complexity computing an optimal allocation is NP-hard even for a single agent (the KNAPSACK problem).

One agent with large weight – find the optimal set of tasks to complete.

Optimal Coalition Structure

Page 8: Overlapping Coalition Formation: Charting  the Tractability Frontier

Theorem: computing an optimal allocation is in P for constant # of agents and poly size weights.

Proof: can be done by dynamic programming.

Optimal Coalition Structure

Page 9: Overlapping Coalition Formation: Charting  the Tractability Frontier

Computational complexity even when weights are at most 3, complex interactions cause NP-hardness (the X3C problem).

Optimal Coalition Structure

Page 10: Overlapping Coalition Formation: Charting  the Tractability Frontier

We assume that:•Weights are polynomially bounded• Interactions are simple.

Optimal Coalition Structure

Page 11: Overlapping Coalition Formation: Charting  the Tractability Frontier

Suppose that the interaction graph is a tree

Optimal Coalition Structure

Page 12: Overlapping Coalition Formation: Charting  the Tractability Frontier

Theorem: if the maximal weight is W and there are n nodes, an optimal allocation can be computed in time linear in n and polynomial in W.

Optimal Coalition Structure

Page 13: Overlapping Coalition Formation: Charting  the Tractability Frontier

We set:ui(xi) – the most an agent can

make working aloneui,j(xi, xj) – the most two agents

can make by working togetherTi(xi) – the most the subtree

rooted at i can make

Optimal Coalition Structure

Page 14: Overlapping Coalition Formation: Charting  the Tractability Frontier

1 8

764

532 9

OPT=max{u1(x1) + §u1,j(x1j,yj) + Tj(wj - yj)}

T3(x3)= max{u3(y3)+§u3,j(y3j,zj) + Tj(wj - zj)}

Page 15: Overlapping Coalition Formation: Charting  the Tractability Frontier

StabilityOptimal resource allocation

Which profit divisions ensure group stability?

Page 16: Overlapping Coalition Formation: Charting  the Tractability Frontier

17,1510,5

1,5

4,310,13

5

5,7

16,5

7 1,1

10,9

4,513,12

(CS, x)CS xOutcome

Is (CS, x) in the core?

Page 17: Overlapping Coalition Formation: Charting  the Tractability Frontier

Deviation“Coalitional game theory [...] considers a game of

n players as a set of possible 2n – 1 coalitions, each of which, call it S, can achieve a particular value v(S) […] against worst case behavior of players in N\S”

C.H. Papadimitriou, STOC 2001

Players assume they are “on their own” if they deviate.

Page 18: Overlapping Coalition Formation: Charting  the Tractability Frontier

17,1510,5

1,5

4,310,13

5

5,7

16,5

7 1,1

10,9

4,513,12

20

15

Page 19: Overlapping Coalition Formation: Charting  the Tractability Frontier

StabilityArbitration functions: agents may receive all or some of the payoff from unbroken/changed agreements.

Behavior can be very general.

Page 20: Overlapping Coalition Formation: Charting  the Tractability Frontier

Arbitration FunctionsOthers can react to deviation

either locally or globally.Conservative – give nothingRefined – give all from unhurt

coalitionsOptimistic – deviators absorb the

marginal damage of deviation; get the difference.

Page 21: Overlapping Coalition Formation: Charting  the Tractability Frontier

17,1510,5

1,5

4,310,13

5

5,7

16,5

7 1,1

10,9

4,513,12

8,15 GlobalLocal

8,10

Page 22: Overlapping Coalition Formation: Charting  the Tractability Frontier

StabilityTheorem: if there is an efficient algorithm to compute the most one can get from global arbitration functions, then P = NP.

Page 23: Overlapping Coalition Formation: Charting  the Tractability Frontier

1

7

6

5

4

32

1 2 3 4 5 6 7

1 2 3 4 5 6 7

05

10

10

10

10

10

10

10

"""

Page 24: Overlapping Coalition Formation: Charting  the Tractability Frontier

StabilityTheorem: if the arbitration function is local, and the interaction graph is a tree,

computing the most a set can get from deviating is possible in poly(n,W) time

Page 25: Overlapping Coalition Formation: Charting  the Tractability Frontier

StabilityDenote the most that a set S can get by deviating by

A*(S,CS, x)Having divided payoffs, can we verify that no set wants to deviate?

Page 26: Overlapping Coalition Formation: Charting  the Tractability Frontier

StabilityTheorem: if the arbitration function is local, and the interaction graph is a tree, then one can verify if an outcome is A -stable in poly(n,W) time.

Page 27: Overlapping Coalition Formation: Charting  the Tractability Frontier

StabilityCorollary: Given a coalition structure CS, we can find x such that (CS, x) is A -stable in poly(n,W) time.

Proof: ellipsoid method to solve an LP

Page 28: Overlapping Coalition Formation: Charting  the Tractability Frontier

RecapOptimization/Stability: Hard in general due to•Weights• Complex interaction

Page 29: Overlapping Coalition Formation: Charting  the Tractability Frontier

More ResultsBounded hyper-treewidth: Our results can be extended to graphs with bounded hyper-treewidth.

If the graph is “tree-like” we can still obtain efficient algorithms.

Page 30: Overlapping Coalition Formation: Charting  the Tractability Frontier

More ResultsStable conservative core: We can find a stable outcome against worst case behavior.

Each agent receives the minimum needed to make his subtree stable.

Page 31: Overlapping Coalition Formation: Charting  the Tractability Frontier

SummaryComputational Issues: A major obstacle in OCF games.

But: if interactions are (somewhat) local, both for values and arbitration functions, we can obtain poly-time algorithms.

Page 32: Overlapping Coalition Formation: Charting  the Tractability Frontier

Poly-time, but…Complexity is still high: Order of O(nkW5(k+1)) for computing optimal allocation in a graph with treewidth k

Can probably do better if valuations are known.

Page 33: Overlapping Coalition Formation: Charting  the Tractability Frontier

Future WorkDeterministic, Exact: randomized/ approximation algorithms?

Restricted classes of games: convex, subadditive…

Page 34: Overlapping Coalition Formation: Charting  the Tractability Frontier

Thank you!

Questions?