Preprocessing Techniques for Computing Nash Equilibria Vincent Conitzer Duke University Based on: Conitzer and Sandholm. A Generalized Strategy Eliminability.

Slides:



Advertisements
Similar presentations
Simple Search Methods for Finding a Nash Equilibrium
Advertisements

Approximate Nash Equilibria in interesting games Constantinos Daskalakis, U.C. Berkeley.
N. Basilico, N. Gatti, F. Amigoni DEI, Politecnico di Milano Leader-Follower Strategies for Robotic Patrolling in Environments with Arbitrary Topology.
S. Ceppi, N. Gatti, and N. Basilico DEI, Politecnico di Milano Computing Bayes-Nash Equilibria through Support Enumeration Methods in Bayesian Two-Player.
Thursday, March 7 Duality 2 – The dual problem, in general – illustrating duality with 2-person 0-sum game theory Handouts: Lecture Notes.
1 Outline relationship among topics secrets LP with upper bounds by Simplex method basic feasible solution (BFS) by Simplex method for bounded variables.
Sampling Distributions
Concise representations of games Vincent Conitzer
Break Time Remaining 10:00.
On Comparing Classifiers : Pitfalls to Avoid and Recommended Approach
CPS Extensive-form games Vincent Conitzer
Joint work with Andre Lieutier Dassault Systemes Domain Theory and Differential Calculus Abbas Edalat Imperial College Oxford.
On the Approximation Performance of Fictitious Play in Finite Games Paul W. GoldbergU. Liverpool Rahul Savani U. Liverpool Troels Bjerre Sørensen U. Warwick.
Graphs, representation, isomorphism, connectivity
Abbas Edalat Imperial College London Contains joint work with Andre Lieutier (AL) and joint work with Marko Krznaric (MK) Data Types.
Faster Query Answering in Probabilistic Databases using Read-Once Functions Sudeepa Roy Joint work with Vittorio Perduca Val Tannen University of Pennsylvania.
Clock will move after 1 minute
Vincent Conitzer CPS Repeated games Vincent Conitzer
Distributed Computing 9. Sorting - a lower bound on bit complexity Shmuel Zaks ©
Monte Hall Problem Let’s Draw a Game Tree… Problem 6, chapter 2.
Complexity Results about Nash Equilibria Vincent Conitzer, Tuomas Sandholm International Joint Conferences on Artificial Intelligence 2003 (IJCAI’03) Presented.
1 A Graph-Theoretic Network Security Game M. Mavronicolas , V. Papadopoulou , A. Philippou  and P. Spirakis § University of Cyprus, Cyprus  University.
EC941 - Game Theory Prof. Francesco Squintani Lecture 4 1.
NON - zero sum games.
Nash’s Theorem Theorem (Nash, 1951): Every finite game (finite number of players, finite number of pure strategies) has at least one mixed-strategy Nash.
Game Theory Assignment For all of these games, P1 chooses between the columns, and P2 chooses between the rows.
COMP 553: Algorithmic Game Theory Fall 2014 Yang Cai Lecture 21.
Computation of Nash Equilibrium Jugal Garg Georgios Piliouras.
Game Theoretical Insights in Strategic Patrolling: Model and Analysis Nicola Gatti – DEI, Politecnico di Milano, Piazza Leonardo.
MIT and James Orlin © Game Theory 2-person 0-sum (or constant sum) game theory 2-person game theory (e.g., prisoner’s dilemma)
Algorithms for solving two- player normal form games Tuomas Sandholm Carnegie Mellon University Computer Science Department.
Part 3: The Minimax Theorem
Extensive-form games. Extensive-form games with perfect information Player 1 Player 2 Player 1 2, 45, 33, 2 1, 00, 5 Players do not move simultaneously.
Vincent Conitzer CPS Normal-form games Vincent Conitzer
by Vincent Conitzer of Duke
Christos alatzidis constantina galbogini.  The Complexity of Computing a Nash Equilibrium  Constantinos Daskalakis  Paul W. Goldberg  Christos H.
Complexity Results about Nash Equilibria
This Wednesday Dan Bryce will be hosting Mike Goodrich from BYU. Mike is going to give a talk during :30 to 12:20 in Main 117 on his work on UAVs.
1 Computing Nash Equilibrium Presenter: Yishay Mansour.
AWESOME: A General Multiagent Learning Algorithm that Converges in Self- Play and Learns a Best Response Against Stationary Opponents Vincent Conitzer.
An Introduction to Game Theory Part III: Strictly Competitive Games Bernhard Nebel.
1 Algorithms for Computing Approximate Nash Equilibria Vangelis Markakis Athens University of Economics and Business.
Segment: Computational game theory Lecture 1b: Complexity Tuomas Sandholm Computer Science Department Carnegie Mellon University.
Techniques for Computing Game-Theoretic Solutions Vincent Conitzer Duke University Parts of this talk are joint work with Tuomas Sandholm.
Simple search methods for finding a Nash equilibrium Ryan Porter, Eugene Nudelman, and Yoav Shoham Games and Economic Behavior, Vol. 63, Issue 2. pp ,
Minimax strategies, Nash equilibria, correlated equilibria Vincent Conitzer
CPS Learning in games Vincent Conitzer
Game representations, solution concepts and complexity Tuomas Sandholm Computer Science Department Carnegie Mellon University.
6.853: Topics in Algorithmic Game Theory Fall 2011 Constantinos Daskalakis Lecture 11.
Standard and Extended Form Games A Lesson in Multiagent System Based on Jose Vidal’s book Fundamentals of Multiagent Systems Henry Hexmoor, SIUC.
Ásbjörn H Kristbjörnsson1 The complexity of Finding Nash Equilibria Ásbjörn H Kristbjörnsson Algorithms, Logic and Complexity.
1 What is Game Theory About? r Analysis of situations where conflict of interests is present r Goal is to prescribe how conflicts can be resolved 2 2 r.
Algorithms for solving two-player normal form games
CPS Utility theory, normal-form games
1 Algorithms for Computing Approximate Nash Equilibria Vangelis Markakis Athens University of Economics and Business.
Parameterized Two-Player Nash Equilibrium Danny Hermelin, Chien-Chung Huang, Stefan Kratsch, and Magnus Wahlstrom..
Commitment to Correlated Strategies Vince Conitzer and Dima Korzhyk.
A useful reduction (SAT -> game)
Non-additive Security Games
A useful reduction (SAT -> game)
Communication Complexity as a Lower Bound for Learning in Games
Commitment to Correlated Strategies
Vincent Conitzer Normal-form games Vincent Conitzer
CPS Extensive-form games
Enumerating All Nash Equilibria for Two-person Extensive Games
Vincent Conitzer Extensive-form games Vincent Conitzer
CPS 173 Extensive-form games
Normal Form (Matrix) Games
A Technique for Reducing Normal Form Games to Compute a Nash Equilibrium Vincent Conitzer and Tuomas Sandholm Carnegie Mellon University, Computer Science.
Presentation transcript:

Preprocessing Techniques for Computing Nash Equilibria Vincent Conitzer Duke University Based on: Conitzer and Sandholm. A Generalized Strategy Eliminability Criterion and Computational Methods for Applying It. AAAI-05 Conitzer and Sandholm. A Technique for Reducing Normal-Form Games to Compute a Nash Equilibrium. AAMAS-06

Computing Nash equilibria in (2-player) normal-form games Computing one Nash equilibrium is PPAD-complete –Daskalakis, Goldberg, & Papadimitriou ECCC05; Chen & Deng FOCS06 Determining whether a Nash equilibrium with a certain property exists is typically NP-complete –Is there an equilibrium with player 1s utility / each players utility / average utility > k? Even hard to approximate –Is there an equilibrium that puts positive / zero probability on pure strategy s? –Etc. –Gilboa & Zemel GEB 89; Conitzer & Sandholm IJCAI03/extended draft All known algorithms take exponential time –Even Lemke-Howson [Savani & von Stengel FOCS04/Econometrica06]

Preprocessing games We are solving for an equilibrium (optimal equilibrium, maybe) When can we shrink the game and solve the shrunk game instead? If a strategy is (strictly) dominated, can throw it out If the game has only a small nonzero component, can focus on that We will see generalizations of both of these

A class of hard games Sandholm, Gilpin, Conitzer AAAI05 0, 20, 33, 00, 00, 20, 00, 2 0, 0 0, 30, 23, 00, 2 2, 42, 0 4, 22, 03, 3 2, 0 2, 42, 04, 2 0, 23, 00, 30, 00, 20, 00, 2 0, 0 3, 00, 20, 30, 2 4, 22, 0 3, 32, 02, 4 1/

Eliminability concepts Dominance: strategy always does worse than some other (mixed) strategy - strong argument - local reasoning - easy to compute - often does not apply Nash equilibrium: strategy does not appear in support of any Nash equilibrium -weaker argument - global reasoning - hard to compute - applies more often 3, 22, 3 3, 2 4, 00, Is there something in between that combines good aspects of both? Yes! [Conitzer & Sandholm AAAI05] 3, 22, 3 3, 2 2, 02, 1.5

Definition as game between attacker and defender Stage 1: Defender specifies probabilities on E strategies (e r * must get > 0) 3, 00, 30, 2 sr4sr4 0, 33, 00, 2 sr3sr3 2, 0 2, 2 sr2sr2 2, 0 2, 2 sr1sr1 sc4sc4 sc3sc3 sc2sc2 sc1sc Stage 2: Attacker chooses one of the E strategies with positive probability to attack and chooses (possibly mixed) attacking strategy attacked attacking Stage 3: Defender chooses on which (non-E) strategy to place the remainder of the probability –If attacking outperforms attacked, attacker wins attacked attacking e r * = s r 3, E r = {s r 3, s r 4 }, E c = {s c 3, s c 4 } 3, 00, 30, 2 sr4sr4 0, 33, 00, 2 sr3sr3 2, 0 2, 2 sr2sr2 2, 0 2, 2 sr1sr1 sc4sc4 sc3sc3 sc2sc2 sc1sc1 3, 00, 30, 2 sr4sr4 0, 33, 00, 2 sr3sr3 2, 0 2, 2 sr2sr2 2, 0 2, 2 sr1sr1 sc4sc4 sc3sc3 sc2sc2 sc1sc1

A spectrum of elimination power The larger the E i sets, the more strategies are eliminable If the E i sets include all strategies, then a strategy is eliminable if and only if no Nash equilibrium places positive probability on it If the E i sets are empty (with the exception of e r *) then e r * is eliminable if and only if it is dominated dominance Nash equilibrium larger E i sets

Alternative definition Stage 1: Defender specifies probabilities on E sets (e r * must get > 0) Stage 2: Attacker chooses one of the E strategies with positive probability to attack Stage 3: Defender distributes the remainder of the probability (not on E) attacked attacked Stage 4: Attacker chooses attacking strategy –If attacking outperforms attacked, attacker wins 0.05 attacked attacking e r * = s r 3, E r = {s r 3, s r 4 }, E c = {s c 3, s c 4 }

Equivalence Theorem. The alternative definition is equivalent to the original one. Proof based on duality (more specifically, Minimax Theorem [von Neumann 1927] )

Mixed integer programming approach (using alternative definition) Continuous variables: p i (e i ), p i e -i (s i ), binary: b i (e i ) maximize p r (e r *) subject to –for both i, for any e i E i, Σp -i (e -i ) + Σp -i e i (s -i ) = 1 –for both i, for any e i E i, p i (e i ) b i (e i ) –for both i, for any e i E i and any d i S i, Σp -i (e -i )(u i (e i, e -i )-u i (d i, e -i )) + Σp -i e i (s -i )(u i (e i, s -i )-u i (d i, s -i )) (b i (e i )-1)U i U i is the maximum difference between two of player is utilities Number of binary variables = |E r | + |E c | –Exponential only in this!

Eliminating strategies in the hard game 0, 20, 33, 00, 00, 20, 00, 2 0, 0 0, 30, 23, 00, 2 2, 42, 0 4, 22, 03, 3 2, 0 2, 42, 04, 2 0, 23, 00, 30, 00, 20, 00, 2 0, 0 3, 00, 20, 30, 2 4, 22, 0 3, 32, 02, 4 1/ ErEr EcEc

Another preprocessing technique for computing a Nash equilibrium [Conitzer & Sandholm AAMAS06] a l, d ml …a 2, d m2 a 1, d m1 ……… a l, d 2l …a 2, d 22 a 1, d 21 a l, d 1l …a 2, d 12 a 1, d 11 c kn, b k …c k2, b k c k1, b k ……… c 2n, b 2 …c 22, b 2 c 21, b 2 c 1n, b 1 …c 12, b 1 c 11, b 1 G π r, π c a l, Σ i p G (s i )d 1l …a 2, Σ i p G (s i )d i2 a 1,Σ i p G (s i )d i1 Σ j p G (t j )c kj, b k … Σ j p G (t j )c 2j, b 2 Σ j p G (t j )c 1j, b 1 G H H

Required structure on original game O a l, d ml …a 2, d m2 a 1, d m1 smsm ………… a l, d 2l …a 2, d 22 a 1, d 21 s2s2 a l, d 1l …a 2, d 12 a 1, d 11 s1s1 c kn, b k …c k2, b k c k1, b k ukuk ………… c 2n, b 2 …c 22, b 2 c 21, b 2 u2u2 c 1n, b 1 …c 12, b 1 c 11, b 1 u1u1 tntn …t2t2 t1t1 vlvl …v2v2 v1v1 That is: against any fixed v j, all the s i give the row player the same utility a j against any fixed u i, all the t j give the column player the same utility b i H G

Solve for equilibrium of G (recursively) smsm … s2s2 s1s1 tntn …t2t2 t1t1 Obtain –Equilibrium distributions p G (s i ), p G (t j ) –Players expected payoffs in equilibrium π r, π c G

Reduced game R π r, π c a l, Σ i p G (s i )d 1l …a 2, Σ i p G (s i )d i2 a 1,Σ i p G (s i )d i1 s Σ j p G (t j )c kj, b k ukuk …… Σ j p G (t j )c 2j, b 2 u2u2 Σ j p G (t j )c 1j, b 1 u1u1 tvlvl …v2v2 v1v1 Expected payoffs when row player plays the equilibrium of G, column player plays v i Expected payoffs when both players play the equilibrium of G Theorem. p R (u i ), p R (s)p G (s i ); p R (v j ), p R (t)p G (t j ) constitutes a Nash equilibrium of original game. H

Example v1v1 t1t1 t2t2 u1u1 2, 20, 32, 3 s1s1 1, 24, 00, 4 s2s2 1, 40, 44, 0 t1t1 t2t2 s1s1 0, 4 s2s2 4, v1v1 t u1u1 2, 21, 3 s 2,

A more difficult example = the game that we solved before! v 1 = b 2 t 1 = b 1 t 2 = b 3 u 1 = a 2 2, 20, 32, 3 s 1 = a 1 1, 24, 00, 4 s 2 = a 3 1, 40, 44, 0 b1b1 b2b2 b3b3 a1a1 1, 20, 4 a2a2 0, 32, 22, 3 a3a3 0, 41, 44, 0 But how (in general) do we find the correct labeling of the strategies as u i, s i, v j, t j ? Can it be done in polynomial time?

Lets try to use satisfiability b1b1 b2b2 b3b3 a1a1 4, 01, 20, 4 a2a2 0, 32, 22, 3 a3a3 0, 41, 44, 0 Say that v(σ) = true if we label σ as one of the s i or t j (that is, we put it in G) If a 1, a 2 are both in G, then b 1 must also be in G because a 1, a 2 get different payoffs against b 1 Equivalently, v(a 1 ) and v(a 2 ) v(b 1 ) –or (-v(a 1 ) or -v(a 2 ) or v(b 1 )) Theorem: satisfaction of all such clauses the condition is satisfied

Clauses for the example b1b1 b2b2 b3b3 a1a1 4, 01, 20, 4 a2a2 0, 32, 22, 3 a3a3 0, 41, 44, 0 v(a 1 ) and v(a 2 ) v(b 1 ) and v(b 2 ) and v(b 3 ) v(a 1 ) and v(a 3 ) v(b 1 ) and v(b 3 ) v(a 2 ) and v(a 3 ) v(b 2 ) and v(b 3 ) v(b 1 ) and v(b 2 ) v(a 1 ) and v(a 2 ) v(b 1 ) and v(b 3 ) v(a 1 ) and v(a 3 ) v(b 2 ) and v(b 3 ) v(a 1 ) and v(a 2 ) and v(a 3 ) Complete characterization of solutions: –Set at most one variable to true for each player (does not reduce game) –Set all variables to true (G = whole game!) –Only nontrivial solution: set v(a 1 ), v(a 3 ), v(b 1 ), v(b 3 ) to true

Simple algorithm Algorithm to find nontrivial solution: –Start with any two variables for the same agent set to true –Follow the implications –If all variables set to true, start with next pair of variables

Solving the example with the algorithm (pass 1) b1b1 b2b2 b3b3 a1a1 4, 01, 20, 4 a2a2 0, 32, 22, 3 a3a3 0, 41, 44, 0 v(a 1 ) and v(a 2 ) v(b 1 ) and v(b 2 ) and v(b 3 ) v(a 1 ) and v(a 3 ) v(b 1 ) and v(b 3 ) v(a 2 ) and v(a 3 ) v(b 2 ) and v(b 3 ) v(b 1 ) and v(b 2 ) v(a 1 ) and v(a 2 ) v(b 1 ) and v(b 3 ) v(a 1 ) and v(a 3 ) v(b 2 ) and v(b 3 ) v(a 1 ) and v(a 2 ) and v(a 3 ) Variables set to true: v(a 1 ) v(a 2 ) v(a 3 ) v(b 1 )v(b 2 )v(b 3 )

Solving the example with the algorithm (pass 2) b1b1 b2b2 b3b3 a1a1 4, 01, 20, 4 a2a2 0, 32, 22, 3 a3a3 0, 41, 44, 0 v(a 1 ) and v(a 2 ) v(b 1 ) and v(b 2 ) and v(b 3 ) v(a 1 ) and v(a 3 ) v(b 1 ) and v(b 3 ) v(a 2 ) and v(a 3 ) v(b 2 ) and v(b 3 ) v(b 1 ) and v(b 2 ) v(a 1 ) and v(a 2 ) v(b 1 ) and v(b 3 ) v(a 1 ) and v(a 3 ) v(b 2 ) and v(b 3 ) v(a 1 ) and v(a 2 ) and v(a 3 ) Variables set to true: v(a 1 ) v(a 3 )v(b 1 )v(b 3 )

Algorithm complexity Theorem. Requires at most O((#rows+#columns) 4 ) clause applications –That is, quadratic if the game is square Can improve in practice by caching previous results

Preprocessing the hard game 2, 44, 23, 3 2, 44, 2 3, 32, 4 0, 21.5, 1.50, 2 2, 42, 04, 23, 3 2, 02, 44, 2 2, 03, 32, 4 0, 20, 33, 00, 00, 20, 00, 2 0, 0 0, 30, 23, 00, 2 2, 42, 0 4, 22, 03, 3 2, 0 2, 42, 04, 2 0, 23, 00, 30, 00, 20, 00, 2 0, 0 3, 00, 20, 30, 2 4, 22, 0 3, 32, 02, 4 0, 33, 0 0, 3 3, 00, 0 1.5, 1.5 3, 00, 30, 0 1.5, 1.50, 0 1.5, 1.5 0, 33, 0 0, 3 3, 00, 0 0, 33, 0 0, 30, 0 3, 00, 3 1/2 1/

Conclusions Generalized strategy eliminability criterion [AAAI05] –Parameterized definition –At one extreme setting, dominance –At other extreme, whether a strategy is in the support of any Nash –Efficiently computable for settings close to dominance Technique for recursively solving subcomponent [AAMAS06] –Subcomponents solution can be used to shrink original game –Efficient algorithm for finding subcomponent Other techniques? Generalization to extensive form? Thank you for your attention!

Search-based approaches Suppose we know the support X i of each player is mixed strategy in equilibrium Then, we have a simple linear feasibility problem: –for both i, for any s i X i, Σp -i (s -i )u i (s i, s -i ) = u i –for both i, for any s i S i - X i, Σp -i (s -i )u i (s i, s -i ) u i Thus, we can search over supports –This is the basic idea underlying methods in Dickhaut & Kaplan 91; Porter, Nudelman, Shoham AAAI04; Sandholm, Gilpin, Conitzer AAAI05 Dominated strategies can be eliminated –A type of preprocessing –What other preprocessing techniques exist?