Correlated Equilibria in Sender-Receiver Games
|
|
- Julian Leonard
- 5 years ago
- Views:
Transcription
1 Correlated Equilibria in Sender-Receiver Games Andreas Blume Department of Economics University of Pittsburgh Pittsburgh, PA May, 2010 Abstract It is shown that the efficiency bound for communication equilibria identified by Goltsman, Hörner, Pavlov and Squintani [13] in the leading example of the Crawford-Sobel model can be obtained with strategy-correlated equilibria. Thus, unlike in earlier implementations of this bound, there is no need for communication to a mediator, for the sender s message to the receiver to be garbled, or for repeated message exchanges between sender and receiver. The occasional mismatch between the encoding and decoding rules used by sender and receiver in a correlated equilibrium can be interpreted as uncertainty about language use. I am grateful to Maria Goltsman and Joel Sobel for comments.
2 1 Introduction The amount of information that can be credibly transmitted from one strategic agent to another is constrained by how closely their preferences are aligned. In their seminal paper on sender-receiver games Crawford and Sobel (CS) [6] study this constraint in a model in which a privately informed sender sends a message to a receiver who then takes an action. Payoffs depend on the sender s private information, referred to as her type, and on the receiver s action but not on the message sent, which is therefore cheap talk. CS find that equilibria have a simple structure. With a type space that is a compact interval, as long as the sender prefers a strictly higher action than the receiver for any realization of the sender s type, in any equilibrium only a finite number of actions is taken and the set of types that induce any given action is a non-degenerate interval. In particular, with fixed preferences equilibria are bounded away from perfect information transmission where the receiver would take a distinct action for every types of the sender. The special case in which the sender s type is uniformly distributed, both sender and receiver have quadratic loss functions and the sender s and receiver s preferred actions differ by a constant bias has received considerable attention. In this uniform-quadratic model equilibria can be ranked in ex-ante efficiency terms and an appealing comparative statics result holds. The equilibrium with the largest number of equilibrium actions is ex-ante efficient and the number of actions in this equilibrium is inversely related to the sender s bias. There are a number of well-known mechanisms for improving efficiency in sender-receiver environments. Myerson [19] gives an example that illustrates the benefits of communicating through a noisy channel; Forges [11] shows that similar effects can be achieved with repeated simultaneous message exchange; Aumann and Hart [2] fully characterize the set of equilibrium payoffs in a class of finite games that are preceded by an infinity of opportunities for simultaneous message exchange; Krishna and Morgan [16] show that two communication rounds suffice to improve on the most efficient equilibrium in the uniform-quadratic CS game for almost all biases, including a range of biases too large to admit any communication in CS; Blume, Board and Kawamura [3] show that for almost any bias in the uniform-quadratic CS game there is a noise level in a noisy-channel game and an equilibrium that strictly improves 1
3 on the best CS equilibrium, again including biases for which communication is impossible in CS; and, Ivanov [14] shows that the same effect can be achieved by replacing the noisy channel with a strategic mediator. It follows from the revelation principle (Myerson [17]) that the set of all equilibrium payoffs that can be attained by one of the above schemes is a subset of the set of payoffs that can be achieved as equilibria of a direct mechanism in which the sender truthfully reports her type to a nonstrategic mediator and the receiver obediently executes the actions recommended to him by the mediator. In general, whenever a Bayesian game is extended by adding a nonstrategic mediator who privately receives messages from the players and privately recommends actions to them, the truthful and obedient equilibria of the extended game are referred to as communication equilibria of the original game (see Forges [9] and Myerson [18]). Recently, Goltsman, Hörner, Pavlov and Squintani (GHPS) [13] derived an efficiency bound for communication equilibria in Crawford-Sobel games with a uniform type distribution, quadratic utility functions and constant bias. The bound can be implemented in Krishna and Morgan s model for a range of biases and in the models of Blume, Board and Kawamura and of Ivanov for all biases. This note shows that the bound can also be implemented without the sender communicating to a nonstrategic mediator, without repeated communication by sender and receiver, without a noisy channel, and without a strategic mediator. Instead, it suffices that there is a correlation device that sends messages to both players before the sender sends her message to the receiver, which is equivalent to saying that the bound can be attained as a (strategy-) correlated equilibrium. Importantly, neither player needs to send messages to the device. The concept of a correlated equilibrium was introduced by Aumann[1] for finite strategic form games. He pointed out that they correspond to the Nash equilibria of games that are obtained by adding a correlation device to the original (base) game that privately sends messages to players before the base game is played. There are multiple ways of extending the correlated equilibrium idea to Bayesian games because of the added possibility that players may send messages to the correlation device before the device sends messages to the players. The most permissive adaptation corresponds to the above mentioned communication 2
4 equilibria. The direct adaptation of Aumann s definition to Bayesian game is the set of strategy-correlated equilibria as defined by Cotter [5]. The set of strategy-correlated equilibria is a (sometimes strict) subset of the set of communication equilibria. Note that since no message is sent to a third party (the correlation device, or a mediator), strategy-correlated equilibria preserve privacy vis-à-vis the third party, as long as the third party cannot observe sender messages and receiver actions. We show that GHPS s efficiency bound for communication equilibria in CS games can be achieved with strategy-correlated equilibria. The proof is constructive. We design a device that probabilistically and privately sends an encoding rule to the sender and a decoding rule to the receiver. The two rules agree (in the sense that the sender can predict the receiver s response, given her message) some but not all of the time. The probability of disagreement is the same as the probability of a transmission error in the noisy channel implementation of the GHPS bound by Blume, Board and Kawamura. One interpretation is that sender and receiver are uncertain about each other s language use. 2 The CS Environment We will be concerned with correlated equilibria in the leading example of Crawford and Sobel [6] (CS). In the uniform-quadratic example of CS a privately informed sender, S, communicates with a receiver, R. Their payoffs depend on the receiver s action, a R, the sender s type t T = [0, 1], and a parameter b > 0 that measures the sender s bias relative to the receiver. The sender s type t is drawn from a uniform distribution on T. When the sender s type is t and the receiver takes action a, the sender s payoff equals U S (a, t, b) = (t + b a) 2 and the receiver s payoff equals U R (a, t) = (t a) 2. CS study the game in which after observing her type t the sender sends a message m M to the receiver, who then takes an action in response to the sender s message. Messages do not directly affect payoffs. We will assume that the message space M equals the unit interval. 1 1 In the CS analysis of their model no assumption about the message space is needed, except that its cardinality be at least as large as the maximal number of equilibrium actions. Our proof makes use of the fact that the message space contains a compact non-degenerate interval. Our assumption that M = [0, 1] meets this requirement. Since we are in essence concerned with designing a communication mechanism, this 3
5 All of this is common knowledge among the players. We will refer to the above payoff structure, type set and type distribution as the (uniform-quadratic) CS environment and once the message and action stages are included as the (uniform-quadratic) CS game. In general communication need not be restricted to the sender sending a single message to the receiver as in the CS game. As discussed earlier, the literature has considered communication through noisy channels, repeated message exchanges and the use of strategic as well as non-strategic mediators, who both receive messages from and send messages to the players. Our focus will be on games in which initially both players privately receive a signal from a correlation device and then the CS game is played, i.e. after both players have received their signals from the device and the sender has learned her type, the sender sends a single message to the receiver who then takes an action. 3 Correlated Equilibria in the CS Game In Bayesian games there are multiple ways of defining correlated equilibria, which vary according to the degree to which correlated signals are allowed to depend on the players types. Strategy-correlated equilibrium, as defined by Cotter [5], can be implemented through a device that sends private signals to players that are independent of the players types. An alternative extension of correlated equilibrium to Bayesian games, that of a communication equilibrium, defined by Forges [9] and Myerson [18], permits players to send messages to the correlation device in addition to receiving instructions from the device. Evidently, the set of communication equilibria is a superset of the set of strategy-correlated equilibria. We will show in this section that in CS games the efficiency bound for communication equilibria can be attained through strategy-correlated equilibria. In CS games a behavior strategy for the sender is a function σ : T (M) and for the receiver it is a function ρ : M R, where we make use of the fact that for any belief the receiver has a unique best action. We write W S (σ, ρ) for the sender s expected payoff and W R (σ, ρ) for the receiver s expected payoff from the strategy profile (σ, ρ). Denote the set of behavior strategies for the sender by Σ and for the receiver by P. A strategy correlated does not appear to be restrictive. 4
6 equilibrium (SCE) in the CS model is a probability distribution λ (Σ P ) that satisfies for all measurable functions δ : Σ Σ and γ : P P : W S (σ, ρ)dλ(σ, ρ) Σ P W R (σ, ρ)dλ(σ, ρ) Σ P W S (δ(σ), ρ)dλ(σ, ρ), and (1) Σ P W R (σ, γ(ρ))dλ(σ, ρ) (2) Σ P An SCE in the CS model can be interpreted as a correlation device that privately sends an encoding rule, σ Σ, to the sender and a decoding rule, ρ P, to the receiver, with the property that both sender and receiver prefer to adhere to their suggested rules. Inequality (1) is the incentive compatibility constraint of sender. It states that the sender must prefer always to use the strategy σ recommended by the device to any deviation rule δ that determines which strategy to follow as a function of the recommendation. Inequality (2) is the analogous constraint for the receiver. Using the revelation principle, Myerson [17], one can characterize the set of communication equilibria in the uniform-quadratic CS model as corresponding to a family of conditional distributions on R, {p( t)} t T, that satisfies: [ ] t = arg max (t + b a) 2 dp(a t ), t T, and (3) t T R ap(a t)dt = tp(a t)dt, a R. (4) T T The family of conditional distributions can be interpreted as a mediator who recommends action a with probability p(a t) to the receiver when the sender reports a type t to the mediator. The first equation (3) is the incentive constraint for the sender. It says that it must be optimal for the sender to report her type truthfully to the mediator. The second equation (4) is the incentive constraint for the receiver and requires that conditional on the action a being recommended, this action is equal to the receiver s expectation of the sender s type conditional on that recommendation; recall that with a quadratic loss function the receiver s optimal action is equal to her expectation of the sender s type. Goltsman, Hörner, 5
7 Pavlov and Squintani (GHPS) [13] use this characterization to show that the receiver s ex ante payoff in any communication equilibrium of the CS model is bounded above by 1b(1 b). Since the ex ante payoffs of the receiver, V 3 R and the sender, V S, are related through V R = V S + b 2, this is also the efficiency bound for communication equilibria CS model. Blume, Board and Kawamura [3] (henceforth BBK) provide a mechanism that attains the efficiency bound for communication equilibria in the CS model. In BBK players communicate through a noisy channel that lets the message pass through with probability ɛ and otherwise transmits a random draw from a distribution G on the message space M = [0, 1] that has an everywhere positive density. 2 BBK show that for any b there exists a noise level ɛ(b) and an equilibrium of the corresponding ɛ(b)-noise game, Γ(ɛ(b)), that achieves the GHPS bound. We will show that this bound can also be attained in a strategy correlated equilibrium of the CS game and therefore requires no communication of the players to a mediator or through a noisy channel. Our proof strategy is to show that for any front-loading equilibrium of the game Γ(ɛ(b)) with noise level ɛ(b) that achieves the GHPS efficiency bound in BBK there exists an outcome-equivalent (and therefore payoff equivalent) strategy-correlated equilibrium. As background it is useful briefly to recall the key elements of the construction of front loading equilibria in BBK. In a front-loading equilibrium the type set, [0, 1], is partitioned into a finite collection of intervals, for any partition element that is not the leftmost interval there is a single message that is sent by types in that partition element and types in the leftmost partition element randomize over all the remaining messages, using the distribution G. The benefit from having the lowest interval send nearly all messages is that upon receiving such messages the receiver assigns relatively high probability to the error event, which biases his action upward toward the mean of the type distribution, which in turn relaxes incentive constraints for the sender. Indeed, using this construction one can show that a sufficiently small amount of noise in message transmission is almost always welfare improving. Moreover, for any b > 0 there exists a noise-level ɛ(b) and a corresponding front-loading equilibrium that achieves the GHPS bound. 2 While the exact nature of the distribution G is irrelevant, it may help to think of it as being the uniform distribution. 6
8 The following result shows that there exists a strategy-correlated equilibrium in the CS game that achieves the GHPS bound. The proof proceeds by constructing a correlation device and an equilibrium for the game that is augmented by the correlation device that translates the BBK-front-loading equilibria of ɛ(b)-noise games into strategy-correlated equilibria. Proposition The CS game with quadratic preferences and a uniform type distribution has a strategy correlated equilibrium that achieves the efficiency bound for communication equilibria. Proof: We will show that any outcome induced by a front-loading equilibrium of an ɛ- noise game can be replicated with a strategy-correlated equilibrium in the game without noise. The result then follows by applying this observation to the case where ɛ = ɛ(b) and the corresponding optimal front-loading equilibrium. Consider any k-step front-loading equilibrium in the ɛ-noise game, Γ(ɛ). In such an equilibrium there is a partition {Θ 1,..., Θ k } = {[0, θ 1 ], (θ 1, θ 2 ],..., (0, θ k ]} of the type space T = [0, 1] such that types in the partition element Θ 1 randomize uniformly over the entire message space M = [0, 1], types in any other partition element Θ i send a distinct message m i M, m i m j i j, i, j > 1, the receiver responds to m i with the action θ i+θ i 1 2 i > 1 and to any other message m m i i > 1 with the action (1 ɛ)θ 1 θ 1 2 +ɛ 1 2 (1 ɛ)θ 1 +ɛ. Note that the latter action is a weighted average of the action 1 2, which would be optimal if the receiver was sure that the message m was generated by noise and of the action θ 1 2, which would be optimal if the receiver was sure that the message m was sent intentionally (by types in the set Θ 1 ); the weights are the posterior probability that message m was received as a result of noise, was received because it was sent intentionally, ɛ (1 ɛ)θ 1, and the posterior probability that it +ɛ (1 ɛ)θ 1 (1 ɛ)θ 1 +ɛ. We will construct a strategy correlated equilibrium in the zero-noise game, Γ(0) that induces the same outcome (joint distribution over types of actions). To this end consider the following correlation device D : The device first draws a pair of potential signals (p 1, p 2 ) from independent uniform distributions on [0, 1] and then privately sends actual signals ξ 1 to the sender and ξ 2 to the receiver. The sender s signal ξ 1 is always equal to p 1. With probability 1 ɛ the receiver s signal ξ 2 equals p 1 as well; otherwise, it equals p 2. Denote the zero-noise 7
9 game augmented by the device D by Γ(0, D). Consider the following strategies in the game Γ(0, D). Sender types in Θ 1 randomize uniformly over M, regardless of the signal ξ 1 that they receive from the correlation device. Types in the interval Θ j, j = 2... k send the message ξ 1 + j 2 ξ k 1 + j 2 k (where x denotes the largest integer less than or equal to x). For j = 2,..., k the receiver responds to message ξ 2 + j 2 ξ k 2 + j 2 θ k with the action j +θ j 1. The receiver s response to any other 2 message is (1 ɛ)θ 1 θ 1 2 +ɛ 1 2 (1 ɛ)θ 1 +ɛ. That these strategies are mutual best replies in Γ(0, D) can be seen as follows: The event that ξ 1 and ξ 2 coincide in Γ(0, D) matches the event that the sender s messages goes through in Γ(ɛ) and has the same probability. If in this event the receiver gets one of the messages of the form ξ 2 + j 2 ξ k 2 + j 2 k, he can with probability one identify the set of types who sent that message, just like in the k-step front-loading equilibrium of Γ(ɛ). If instead in this event he receives any of the other messages, he does not know whether this is because of randomization by a type in the lowest partition element or because the signals ξ 1 and ξ 2 differ. Therefore, in this case his best reply will be a weighted average that assigns probability (1 ɛ)θ 1 (1 ɛ)θ 1 +ɛ Θ 1, and probability to the event that the message was sent by a type in the lowest interval, ɛ (1 ɛ)θ 1 +ɛ to the event that it was sent by one of the other types, again just like in the k-step front-loading equilibrium of Γ(ɛ). In the event that ξ 1 and ξ 2 do not coincide the receivers s response is the same weighted average and agrees with his action in the k-step front-loading equilibrium of Γ(ɛ). Finally, the incentives of the sender are the same in the games Γ(ɛ) and Γ(0, D). In one game, Γ(0, D), the sender can only affect the receiver s action with positive probability in the event that the signals ξ 1 and ξ 2 coincide, and in the other, Γ(ɛ), she can only affect the receiver s action with positive probability in the no noise event, in which the message is transmitted as sent. Thus in both cases the arbitrage condition that makes the boundary type θ i indifferent between the action that is induced by types in Θ i and the one induced by types in Θ i+1 is the same as in the CS model for all i = 1,..., k. 8
10 4 Additional Comments on the Literature As noted earlier, one interpretation of a correlated equilibrium in the CS game is as describing situations where sender and receiver are uncertain about each other s language use. De Jaegher [7] has linked strategic vagueness to correlated equilibrium and illustrated the potential benefits of such vagueness by example. 3 Our result implements this idea in CS games. A number of authors have explored the relation between communication equilibria and correlated equilibria in Bayesian games. Forges [8] studies infinitely repeated finite senderreceiver games with limit of the average payoffs and finds that the sets of communication equilibrium payoffs of such games coincide with the sets of correlated equilibrium payoffs. Forges [12] shows that correlation followed by two rounds of interim communication can achieve all communication equilibrium payoffs in finite games with at least three players, provided the message space has the cardinality of the continuum. Vida [20] shows that the set of communication equilibrium payoffs of any finite Bayesian game can be realized via correlated equilibria of an extended game in which players have infinitely many opportunities to exchange messages from a countable message space before they play the base game. The present paper, in contrast, is concerned with comparing the efficiency frontiers of communication equilibria and correlated equilibria in one-shot CS games, i.e. with one round of interim communication. Most directly related to the present paper is Forges [10], which implies the equivalence of the sets of correlated and communication equilibrium payoffs in finite sender-receiver games. Because of the finiteness requirement, this result does not directly carry over to CS games with their continuous type and action spaces. In addition, finiteness is relied upon in the proof; e.g., the encryption of the sender s strategy involves a random draw from a uniform distribution over the set of all permutations of the outcome space (the set of type-action pairs). A natural open question is how to extend Forges [10] result to a class of infinite games that includes CS games. 3 In his paper, as in Blume and Board [4], the focus is on the strategic use of an existing vague language, as opposed to the question of why language is vague that is raised by Lipman [15]. 9
11 References [1] Aumann, Robert J. [1974], Subjectivity and Correlation in Randomized Strategies, Journal of Mathematical Economics 1, [2] Aumann, Robert J. and Sergiu Hart [2003], Long Cheap Talk, Econometrica 71, [3] Blume, Andreas, Oliver J. Board and Kohei Kawamura [2007]: Noisy Talk, Theoretical Economics 2, [4] Blume, Andreas and Oliver J. Board [2009], Intentional Vagueness, University of Pittsburgh working paper. [5] Cotter, Kevin D. [1991], Correlated Equilibrium in Games with Type-Dependent Strategies, Journal of Economic Theory 54, [6] Crawford, V.P.and J. Sobel [1982], Strategic Information Transmission, Econometrica 50, [7] De Jaegher, Kris [2003], A Game-Theoretic Rationale for Vagueness, Linguistics and Philosophy 26, [8] Forges, Françoise [1985], Correlated Equilibria in a Class of Repeated Games with Incomplete Information, International Journal of Game Theory 14, [9] Forges, Françoise [1986], An Approach to Communication Equilibria, Econometrica 54, [10] Forges, Françoise [1988], Can Sunspots Replace a Mediator, Journal of Mathematical Economics 17, [11] Forges, Françoise [1990], Equilibria with Communication in a Job Market Example, Quarterly Journal of Economics 105, [12] Forges, Françoise [1990], Universal Mechanisms, Econometrica 58,
12 [13] Goltsman, Maria, Johannes Hörner, Gregory Pavlov, and Francesco Squintani [2009], Mediation, Arbitration and Negotiation, Journal of Economic Theory 144, [14] Ivanov, Maxim [2009], Communication via a Strategic Mediator, Journal of Economic Theory, forthcoming. [15] Lipman, Barton L. [2009], Why is Language Vague?, Boston University working paper. [16] Krishna, Vijay and John Morgan [2004], The Art of Conversation: Eliciting Information from Experts Through Multi-Stage Communication, Journal of Economic Theory 117, [17] Myerson, Roger B. [1982], Optimal Coordination Mechanisms in Generalized Principal-Agent Problems, Journal of Mathematical Economics 10, [18] Myerson, Roger B. [1986], Multistage Games with Communication, Econometrica 54, [19] Myerson, Roger B. [1991], Game Theory: Analysis of Conflict, Harvard University Press, Cambridge, MA. [20] Vida, Péter [2007], From Communication Equilibria to Correlated Equilibria, University of Vienna working paper. 11
Language Barriers. Andreas Blume and Oliver Board Department of Economics University of Pittsburgh Pittsburgh, PA October 5, 2009.
Language Barriers Andreas Blume and Oliver Board Department of Economics University of Pittsburgh Pittsburgh, PA 15260 October 5, 2009 Abstract Private information about language competence drives a wedge
More informationFailure of Common Knowledge of Language in Common-Interest Communication Games
Failure of Common Knowledge of Language in Common-Interest Communication Games Andreas Blume Department of Economics University of Arizona May, 206 Abstract This paper explores the fault line that separates
More informationExtensive Games with Imperfect Information
Extensive Games with Imperfect Information Definition (Os 314.1): An extensive game with imperfect information consists of a set of players N a set of terminal histories H; no sequence is a proper subhistory
More informationOvercommunication in Strategic Information Transmission Games
Overcommunication in Strategic Information Transmission Games Hongbin Cai, Joseph Tao-Yi Wang Department of Economics, UCLA, Los Angeles, CA 90095, USA Abstract Since Crawford and Sobel (1982), the theory
More informationMechanism Design in Large Congestion Games
Mechanism Design in Large Congestion Games Ryan Rogers, Aaron Roth, Jonathan Ullman, and Steven Wu July 22, 2015 Routing Game l e (y) Routing Game Routing Game A routing game G is defined by Routing Game
More informationDynamic Information Revelation in Cheap Talk
Dynamic Information Revelation in Cheap Talk Maxim Ivanov y McMaster University September 4 Abstract This paper studies a multi-stage version of Crawford Sobel s (98) communication game. In every period
More informationCommunication in a Complicated World
Communication in a Complicated World Steve Callander Nicolas S. Lambert Niko Matouschek September 20, 2017 PRELIMINARY AND INCOMPLETE Abstract Experts play a central role in many economic, political, and
More informationLying and Deception in Games
Lying and Deception in Games Joel Sobel July 6, 2016 Abstract This article proposes definitions of lying and deception in strategic settings. A critical distinction is that deception requires a model of
More informationNetwork Topology and Equilibrium Existence in Weighted Network Congestion Games
Network Topology and Equilibrium Existence in Weighted Network Congestion Games Igal Milchtaich, Bar-Ilan University August 2010 Abstract. Every finite noncooperative game can be presented as a weighted
More informationKNOWLEDGE TRANSFER AS A RATIONAL CHOICE: A DECISION THEORETIC CHARACTERIZATION
NOWLEDGE TRANSFER AS A RATIONAL CHOICE: A DECISION THEORETIC CHARACTERIZATION Yasuo Sasaki School of nowledge Science, Japan Advanced Institute of Science and Technology 1-1, Asahidai, Nomi, Ishikawa,
More informationLying and Deception in Games
Lying and Deception in Games Joel Sobel February 14, 2018 Abstract This article proposes definitions of lying and deception in strategic settings. A critical distinction is that deception requires a model
More informationAn Extension on the Standard Cheap-Talk Model of Crawford and Sobel: The Sender First Needs to Collect Information About his Type
ERASMUS UIVERSITY ROTTERDAM ERASMUS SCHOOL OF ECOOMICS DEPARTMET OF ECOOMICS Master Thesis Master of Economics & Business MSc Economics of Management and Organizations June 9, 05 An Extension on the Standard
More informationComment on Strategic Information Management Under Leakage in a. Supply Chain
Comment on Strategic Information Management Under Leakage in a Supply Chain Lin Tian 1 School of International Business Administration, Shanghai University of Finance and Economics, 00433 Shanghai, China,
More informationEquilibrium Tracing in Bimatrix Games
Equilibrium Tracing in Bimatrix Games Anne Balthasar Department of Mathematics, London School of Economics, Houghton St, London WCA AE, United Kingdom A.V.Balthasar@lse.ac.uk Abstract. We analyze the relations
More informationapproach is to rely on individuals in a system to select tasks themselves. Both approaches
Chapter 5 Task Routing Engaging a crowd to tackle complex tasks relies not only on effective coordination, but on recruiting individuals with relevant expertise to join the problem-solving effort. One
More informationUncertainty Regarding Interpretation of the `Negligence Rule' and Its Implications for the Efficiency of Outcomes
Jawaharlal Nehru University From the SelectedWorks of Satish K. Jain 2011 Uncertainty Regarding Interpretation of the `Negligence Rule' and Its Implications for the Efficiency of Outcomes Satish K. Jain,
More informationStrategic Listening. Paul E. Fischer Mirko S. Heinle Kevin Smith. University of Pennsylvania. August 2016
Strategic Listening Paul E. Fischer Mirko S. Heinle Kevin Smith University of Pennsylvania August 2016 Abstract: We consider a cheap talk setting with two senders and a continuum of receivers with heterogenous
More informationOn the Computational Complexity of Nash Equilibria for (0, 1) Bimatrix Games
On the Computational Complexity of Nash Equilibria for (0, 1) Bimatrix Games Bruno Codenotti Daniel Štefankovič Abstract The computational complexity of finding a Nash equilibrium in a nonzero sum bimatrix
More informationRules and Commitment in Communication
Rules and Commitment in Communication Guillaume Fréchette Alessandro Lizzeri Jacopo Perego New York University New York University Yale University October 23, 2017 Preliminary and Incomplete. Abstract
More informationCompetition and Sequential Communication: An Experimental Study
Competition and Sequential Communication: An Experimental Study William Minozzi Jonathan Woon February 9, 2017 Abstract We experimentally investigate Krishna and Morgan s (2001b) model of expertise, varying
More informationExtremal Graph Theory: Turán s Theorem
Bridgewater State University Virtual Commons - Bridgewater State University Honors Program Theses and Projects Undergraduate Honors Program 5-9-07 Extremal Graph Theory: Turán s Theorem Vincent Vascimini
More informationWait and See. Péter Eső and Yuk-fai Fong Kellogg School of Management Northwestern University. November 2007
Wait and See Péter Eső and Yuk-fai Fong Kellogg School of Management Northwestern University November 2007 Abstract We study a dynamic cheap talk model with multiple senders where the receiver can choose
More informationMA651 Topology. Lecture 4. Topological spaces 2
MA651 Topology. Lecture 4. Topological spaces 2 This text is based on the following books: Linear Algebra and Analysis by Marc Zamansky Topology by James Dugundgji Fundamental concepts of topology by Peter
More informationHow Cheap Talk Enhances Efficiency in Public Goods Games
How Cheap Talk Enhances Efficiency in Public Goods Games Thomas Palfrey Howard Rosenthal Nilanjan Roy January 12, 215 Abstract This paper uses a Bayesian mechanism design approach to investigate the effects
More informationRules and Commitment in Communication
Rules and Commitment in Communication Guillaume Fréchette Alessandro Lizzeri Jacopo Perego NYU Harvard October 2017 INTRODUCTION We revisit a classic question in economics from a new perspective: How much
More informationHowever, this is not always true! For example, this fails if both A and B are closed and unbounded (find an example).
98 CHAPTER 3. PROPERTIES OF CONVEX SETS: A GLIMPSE 3.2 Separation Theorems It seems intuitively rather obvious that if A and B are two nonempty disjoint convex sets in A 2, then there is a line, H, separating
More informationCAP 5993/CAP 4993 Game Theory. Instructor: Sam Ganzfried
CAP 5993/CAP 4993 Game Theory Instructor: Sam Ganzfried sganzfri@cis.fiu.edu 1 Announcements HW 1 due today HW 2 out this week (2/2), due 2/14 2 Definition: A two-player game is a zero-sum game if for
More informationRules and Commitment in Communication
Rules and Commitment in Communication Guillaume Fréchette Alessandro Lizzeri Jacopo Perego New York University New York University Yale University January 31, 2018 Preliminary and Incomplete. Abstract
More informationCompetition, Preference Uncertainty, and Jamming: A Strategic Communication Experiment. William Minozzi Jonathan Woon.
Competition, Preference Uncertainty, and Jamming: A Strategic Communication Experiment William Minozzi Jonathan Woon March 10, 2014 Abstract We conduct a game-theoretic laboratory experiment to investigate
More informationA Dynamic Model of Network Formation
Games and Economic Behavior 34, 331 341 Ž 001. doi:10.1006 game.000.0803, available online at http: www.idealibrary.com on A Dynamic Model of Network Formation Alison Watts* Department of Economics, Box
More informationSequential Equilibrium in Games of Imperfect Recall
Sequential Equilibrium in Games of Imperfect Recall Joseph Y. Halpern Cornell University halpern@cs.cornell.edu Rafael Pass Cornell University rafael@cs.cornell.edu First version: October, 2008 This version:
More informationAlgorithmic Game Theory and Applications. Lecture 16: Selfish Network Routing, Congestion Games, and the Price of Anarchy
Algorithmic Game Theory and Applications Lecture 16: Selfish Network Routing, Congestion Games, and the Price of Anarchy Kousha Etessami warning, again 1 In the few remaining lectures, we will briefly
More information2 The Fractional Chromatic Gap
C 1 11 2 The Fractional Chromatic Gap As previously noted, for any finite graph. This result follows from the strong duality of linear programs. Since there is no such duality result for infinite linear
More informationMIDTERM EXAMINATION Networked Life (NETS 112) November 21, 2013 Prof. Michael Kearns
MIDTERM EXAMINATION Networked Life (NETS 112) November 21, 2013 Prof. Michael Kearns This is a closed-book exam. You should have no material on your desk other than the exam itself and a pencil or pen.
More information6. Lecture notes on matroid intersection
Massachusetts Institute of Technology 18.453: Combinatorial Optimization Michel X. Goemans May 2, 2017 6. Lecture notes on matroid intersection One nice feature about matroids is that a simple greedy algorithm
More informationAlgorithmic Game Theory and Applications. Lecture 16: Selfish Network Routing, Congestion Games, and the Price of Anarchy.
Algorithmic Game Theory and Applications Lecture 16: Selfish Network Routing, Congestion Games, and the Price of Anarchy Kousha Etessami games and the internet Basic idea: The internet is a huge experiment
More informationRepresentation of Finite Games as Network Congestion Games
Representation of Finite Games as Network Congestion Games Igal Milchtaich To cite this version: Igal Milchtaich. Representation of Finite Games as Network Congestion Games. Roberto Cominetti and Sylvain
More informationGeneral properties of staircase and convex dual feasible functions
General properties of staircase and convex dual feasible functions JÜRGEN RIETZ, CLÁUDIO ALVES, J. M. VALÉRIO de CARVALHO Centro de Investigação Algoritmi da Universidade do Minho, Escola de Engenharia
More informationTOPOLOGY, DR. BLOCK, FALL 2015, NOTES, PART 3.
TOPOLOGY, DR. BLOCK, FALL 2015, NOTES, PART 3. 301. Definition. Let m be a positive integer, and let X be a set. An m-tuple of elements of X is a function x : {1,..., m} X. We sometimes use x i instead
More informationROUGH MEMBERSHIP FUNCTIONS: A TOOL FOR REASONING WITH UNCERTAINTY
ALGEBRAIC METHODS IN LOGIC AND IN COMPUTER SCIENCE BANACH CENTER PUBLICATIONS, VOLUME 28 INSTITUTE OF MATHEMATICS POLISH ACADEMY OF SCIENCES WARSZAWA 1993 ROUGH MEMBERSHIP FUNCTIONS: A TOOL FOR REASONING
More informationThis article was originally published in a journal published by Elsevier, and the attached copy is provided by Elsevier for the author s benefit and for the benefit of the author s institution, for non-commercial
More information6 Extensive Form Games
6 Extensive Form Games 6.1 Example: Representing a Simultaneous 22 Game Alice H HHHH O H HH o Q Bob H QQQ h o HHHH h 2 1 1 2 Figure 1: The Battle of the Sexes in Extensive Form So far we have described
More informationOn the Efficiency of Negligence Rule
Jawaharlal Nehru University From the SelectedWorks of Satish K. Jain 2009 On the Efficiency of Negligence Rule Satish K. Jain, Jawaharlal Nehru University Available at: https://works.bepress.com/satish_jain/2/
More informationPathways to Equilibria, Pretty Pictures and Diagrams (PPAD)
1 Pathways to Equilibria, Pretty Pictures and Diagrams (PPAD) Bernhard von Stengel partly joint work with: Marta Casetti, Julian Merschen, Lászlo Végh Department of Mathematics London School of Economics
More informationA GRAPH FROM THE VIEWPOINT OF ALGEBRAIC TOPOLOGY
A GRAPH FROM THE VIEWPOINT OF ALGEBRAIC TOPOLOGY KARL L. STRATOS Abstract. The conventional method of describing a graph as a pair (V, E), where V and E repectively denote the sets of vertices and edges,
More informationEdge-exchangeable graphs and sparsity
Edge-exchangeable graphs and sparsity Tamara Broderick Department of EECS Massachusetts Institute of Technology tbroderick@csail.mit.edu Diana Cai Department of Statistics University of Chicago dcai@uchicago.edu
More informationTHE preceding chapters were all devoted to the analysis of images and signals which
Chapter 5 Segmentation of Color, Texture, and Orientation Images THE preceding chapters were all devoted to the analysis of images and signals which take values in IR. It is often necessary, however, to
More informationUncertain Data Models
Uncertain Data Models Christoph Koch EPFL Dan Olteanu University of Oxford SYNOMYMS data models for incomplete information, probabilistic data models, representation systems DEFINITION An uncertain data
More informationThe Price of Selfishness in Network Coding Jason R. Marden, Member, IEEE, and Michelle Effros, Fellow, IEEE
IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 58, NO. 4, APRIL 2012 2349 The Price of Selfishness in Network Coding Jason R. Marden, Member, IEEE, and Michelle Effros, Fellow, IEEE Abstract A game-theoretic
More information1.1 What is Microeconomics?
1.1 What is Microeconomics? Economics is the study of allocating limited resources to satisfy unlimited wants. Such a tension implies tradeoffs among competing goals. The analysis can be carried out at
More informationEfficient Random Assignment with Constrained Rankings
Efficient Random Assignment with Constrained Rankings Gabriel Carroll Department of Economics, Massachusetts Institute of Technology E52-391, 50 Memorial Drive, Cambridge MA 02142, USA gdc@mit.edu January
More informationLecture notes on the simplex method September We will present an algorithm to solve linear programs of the form. maximize.
Cornell University, Fall 2017 CS 6820: Algorithms Lecture notes on the simplex method September 2017 1 The Simplex Method We will present an algorithm to solve linear programs of the form maximize subject
More informationLecture 2 September 3
EE 381V: Large Scale Optimization Fall 2012 Lecture 2 September 3 Lecturer: Caramanis & Sanghavi Scribe: Hongbo Si, Qiaoyang Ye 2.1 Overview of the last Lecture The focus of the last lecture was to give
More informationConvexization in Markov Chain Monte Carlo
in Markov Chain Monte Carlo 1 IBM T. J. Watson Yorktown Heights, NY 2 Department of Aerospace Engineering Technion, Israel August 23, 2011 Problem Statement MCMC processes in general are governed by non
More informationConstraint Satisfaction Algorithms for Graphical Games
Constraint Satisfaction Algorithms for Graphical Games Vishal Soni soniv@umich.edu Satinder Singh baveja@umich.edu Computer Science and Engineering Division University of Michigan, Ann Arbor Michael P.
More informationarxiv: v1 [math.co] 28 Sep 2010
Densities of Minor-Closed Graph Families David Eppstein Computer Science Department University of California, Irvine Irvine, California, USA arxiv:1009.5633v1 [math.co] 28 Sep 2010 September 3, 2018 Abstract
More informationSolution of Rectangular Interval Games Using Graphical Method
Tamsui Oxford Journal of Mathematical Sciences 22(1 (2006 95-115 Aletheia University Solution of Rectangular Interval Games Using Graphical Method Prasun Kumar Nayak and Madhumangal Pal Department of Applied
More informationMultiple Agents. Why can t we all just get along? (Rodney King) CS 3793/5233 Artificial Intelligence Multiple Agents 1
Multiple Agents Why can t we all just get along? (Rodney King) CS 3793/5233 Artificial Intelligence Multiple Agents 1 Assumptions Assumptions Definitions Partially bservable Each agent can act autonomously.
More informationThe Structure of Bull-Free Perfect Graphs
The Structure of Bull-Free Perfect Graphs Maria Chudnovsky and Irena Penev Columbia University, New York, NY 10027 USA May 18, 2012 Abstract The bull is a graph consisting of a triangle and two vertex-disjoint
More information1. Lecture notes on bipartite matching February 4th,
1. Lecture notes on bipartite matching February 4th, 2015 6 1.1.1 Hall s Theorem Hall s theorem gives a necessary and sufficient condition for a bipartite graph to have a matching which saturates (or matches)
More informationPart II. Graph Theory. Year
Part II Year 2017 2016 2015 2014 2013 2012 2011 2010 2009 2008 2007 2006 2005 2017 53 Paper 3, Section II 15H Define the Ramsey numbers R(s, t) for integers s, t 2. Show that R(s, t) exists for all s,
More informationIntroduction to Algorithms / Algorithms I Lecturer: Michael Dinitz Topic: Algorithms and Game Theory Date: 12/3/15
600.363 Introduction to Algorithms / 600.463 Algorithms I Lecturer: Michael Dinitz Topic: Algorithms and Game Theory Date: 12/3/15 25.1 Introduction Today we re going to spend some time discussing game
More informationProgress Towards the Total Domination Game 3 4 -Conjecture
Progress Towards the Total Domination Game 3 4 -Conjecture 1 Michael A. Henning and 2 Douglas F. Rall 1 Department of Pure and Applied Mathematics University of Johannesburg Auckland Park, 2006 South Africa
More informationRandom Oracles - OAEP
Random Oracles - OAEP Anatoliy Gliberman, Dmitry Zontov, Patrick Nordahl September 23, 2004 Reading Overview There are two papers presented this week. The first paper, Random Oracles are Practical: A Paradigm
More information1 Linear programming relaxation
Cornell University, Fall 2010 CS 6820: Algorithms Lecture notes: Primal-dual min-cost bipartite matching August 27 30 1 Linear programming relaxation Recall that in the bipartite minimum-cost perfect matching
More informationCOMPRESSED SENSING: A NOVEL POLYNOMIAL COMPLEXITY SOLUTION TO NASH EQUILIBRIA IN DYNAMICAL GAMES. Jing Huang, Liming Wang and Dan Schonfeld
COMPRESSED SENSING: A NOVEL POLYNOMIAL COMPLEXITY SOLUTION TO NASH EQUILIBRIA IN DYNAMICAL GAMES Jing Huang, Liming Wang and Dan Schonfeld Department of Electrical and Computer Engineering, University
More informationA Little Point Set Topology
A Little Point Set Topology A topological space is a generalization of a metric space that allows one to talk about limits, convergence, continuity and so on without requiring the concept of a distance
More informationARELAY network consists of a pair of source and destination
158 IEEE TRANSACTIONS ON INFORMATION THEORY, VOL 55, NO 1, JANUARY 2009 Parity Forwarding for Multiple-Relay Networks Peyman Razaghi, Student Member, IEEE, Wei Yu, Senior Member, IEEE Abstract This paper
More information1 A Tale of Two Lovers
CS 120/ E-177: Introduction to Cryptography Salil Vadhan and Alon Rosen Dec. 12, 2006 Lecture Notes 19 (expanded): Secure Two-Party Computation Recommended Reading. Goldreich Volume II 7.2.2, 7.3.2, 7.3.3.
More informationAbstract Combinatorial Games
Abstract Combinatorial Games Arthur Holshouser 3600 Bullard St. Charlotte, NC, USA Harold Reiter Department of Mathematics, University of North Carolina Charlotte, Charlotte, NC 28223, USA hbreiter@email.uncc.edu
More informationSequential Two-Prize Contests
Sequential Two-Prize Contests Aner Sela September, 2009 Abstract We study two-stage all-pay auctions with two identical prizes. In each stage, the players compete for one prize. Each player may win either
More informationTreewidth and graph minors
Treewidth and graph minors Lectures 9 and 10, December 29, 2011, January 5, 2012 We shall touch upon the theory of Graph Minors by Robertson and Seymour. This theory gives a very general condition under
More informationLecture 10, Zero Knowledge Proofs, Secure Computation
CS 4501-6501 Topics in Cryptography 30 Mar 2018 Lecture 10, Zero Knowledge Proofs, Secure Computation Lecturer: Mahmoody Scribe: Bella Vice-Van Heyde, Derrick Blakely, Bobby Andris 1 Introduction Last
More informationADENSE wireless network is defined to be a wireless network
1094 IEEE/ACM TRANSACTIONS ON NETWORKING, VOL 16, NO 5, OCTOBER 2008 The Analysis of Nash Equilibria of the One-Shot Random-Access Game for Wireless Networks and the Behavior of Selfish Nodes Hazer Inaltekin,
More informationCHAPTER 13: FORMING COALITIONS. Multiagent Systems. mjw/pubs/imas/
CHAPTER 13: FORMING COALITIONS Multiagent Systems http://www.csc.liv.ac.uk/ mjw/pubs/imas/ Coalitional Games Coalitional games model scenarios where agents can benefit by cooperating. Issues in coalitional
More informationOn Approximating Minimum Vertex Cover for Graphs with Perfect Matching
On Approximating Minimum Vertex Cover for Graphs with Perfect Matching Jianer Chen and Iyad A. Kanj Abstract It has been a challenging open problem whether there is a polynomial time approximation algorithm
More informationOn the Max Coloring Problem
On the Max Coloring Problem Leah Epstein Asaf Levin May 22, 2010 Abstract We consider max coloring on hereditary graph classes. The problem is defined as follows. Given a graph G = (V, E) and positive
More informationOptimal Routing Control: Repeated Game Approach
IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 47, NO. 3, MARCH 2002 437 Optimal Routing Control: Repeated Game Approach Richard J. La and Venkat Anantharam, Fellow, IEEE Abstract Communication networks
More informationM3P1/M4P1 (2005) Dr M Ruzhansky Metric and Topological Spaces Summary of the course: definitions, examples, statements.
M3P1/M4P1 (2005) Dr M Ruzhansky Metric and Topological Spaces Summary of the course: definitions, examples, statements. Chapter 1: Metric spaces and convergence. (1.1) Recall the standard distance function
More informationTopological properties of convex sets
Division of the Humanities and Social Sciences Ec 181 KC Border Convex Analysis and Economic Theory Winter 2018 Topic 5: Topological properties of convex sets 5.1 Interior and closure of convex sets Let
More informationThe Game Chromatic Number of Some Classes of Graphs
The Game Chromatic Number of Some Classes of Graphs Casper Joseph Destacamento 1, Andre Dominic Rodriguez 1 and Leonor Aquino-Ruivivar 1,* 1 Mathematics Department, De La Salle University *leonorruivivar@dlsueduph
More information2017 SOLUTIONS (PRELIMINARY VERSION)
SIMON MARAIS MATHEMATICS COMPETITION 07 SOLUTIONS (PRELIMINARY VERSION) This document will be updated to include alternative solutions provided by contestants, after the competition has been mared. Problem
More informationUsing Arithmetic of Real Numbers to Explore Limits and Continuity
Using Arithmetic of Real Numbers to Explore Limits and Continuity by Maria Terrell Cornell University Problem Let a =.898989... and b =.000000... (a) Find a + b. (b) Use your ideas about how to add a and
More informationHedonic Clustering Games
Hedonic Clustering Games [Extended Abstract] Moran Feldman CS Dept., Technion Haifa, Israel moranfe@cs.technion.ac.il Liane Lewin-Eytan IBM Haifa Research Lab. Haifa, Israel lianel@il.ibm.com Joseph (Seffi)
More informationA Search Theoretical Approach to P2P Networks: Analysis of Learning
A Search Theoretical Approach to P2P Networks: Analysis of Learning Nazif Cihan Taş Dept. of Computer Science University of Maryland College Park, MD 2742 Email: ctas@cs.umd.edu Bedri Kâmil Onur Taş Dept.
More informationHow do networks form? Strategic network formation
How do networks form? Strategic network formation Mihaela van der Schaar University of California, Los Angeles Acknowledgement: ONR 1 Social networks Friendship networks Work networks Scientific networks
More informationXI International PhD Workshop OWD 2009, October Fuzzy Sets as Metasets
XI International PhD Workshop OWD 2009, 17 20 October 2009 Fuzzy Sets as Metasets Bartłomiej Starosta, Polsko-Japońska WyŜsza Szkoła Technik Komputerowych (24.01.2008, prof. Witold Kosiński, Polsko-Japońska
More informationAn Eternal Domination Problem in Grids
Theory and Applications of Graphs Volume Issue 1 Article 2 2017 An Eternal Domination Problem in Grids William Klostermeyer University of North Florida, klostermeyer@hotmail.com Margaret-Ellen Messinger
More informationMTAEA Convexity and Quasiconvexity
School of Economics, Australian National University February 19, 2010 Convex Combinations and Convex Sets. Definition. Given any finite collection of points x 1,..., x m R n, a point z R n is said to be
More informationCommunication Amid Uncertainty
Communication Amid Uncertainty Madhu Sudan Microsoft Research Based on Juba, S. (STOC 2008, ITCS 2011) Goldreich, Juba, S. (JACM 2011) Juba, Kalai, Khanna, S. (ITCS 2011) Haramaty, S. (ITCS 2014) Canonne,
More informationUnit 1 Algebraic Functions and Graphs
Algebra 2 Unit 1 Algebraic Functions and Graphs Name: Unit 1 Day 1: Function Notation Today we are: Using Function Notation We are successful when: We can Use function notation to evaluate a function This
More informationBayesian Action-Graph Games
Bayesian Action-Graph Games Albert Xin Jiang and Kevin Leyton-Brown Department of Computer Science University of British Columbia November 13, 2011 Equilibrium Computation in Bayesian Games Equilibrium
More informationCombinatorial Auctions: A Survey by de Vries and Vohra
Combinatorial Auctions: A Survey by de Vries and Vohra Ashwin Ganesan EE228, Fall 2003 September 30, 2003 1 Combinatorial Auctions Problem N is the set of bidders, M is the set of objects b j (S) is the
More informationTopology - I. Michael Shulman WOMP 2004
Topology - I Michael Shulman WOMP 2004 1 Topological Spaces There are many different ways to define a topological space; the most common one is as follows: Definition 1.1 A topological space (often just
More informationMathematical and Algorithmic Foundations Linear Programming and Matchings
Adavnced Algorithms Lectures Mathematical and Algorithmic Foundations Linear Programming and Matchings Paul G. Spirakis Department of Computer Science University of Patras and Liverpool Paul G. Spirakis
More informationUnifying and extending hybrid tractable classes of CSPs
Journal of Experimental & Theoretical Artificial Intelligence Vol. 00, No. 00, Month-Month 200x, 1 16 Unifying and extending hybrid tractable classes of CSPs Wady Naanaa Faculty of sciences, University
More informationChoice under Social Constraints
Choice under Social Constraints T.C.A Madhav Raghavan January 28, 2011 Abstract Agents choose a best element from the available set. But the available set may itself depend on the choices of other agents.
More informationNetwork Selection and Handoff in Wireless Networks: A Game Theoretic Approach
Chapter 23 Network Selection and Handoff in Wireless Networks: A Game Theoretic Approach Josephina Antoniou, Vicky Papadopoulou, Vasos Vassiliou, and Andreas Pitsillides Contents 23.1 Introduction... 536
More informationUsing HMM in Strategic Games
Using HMM in Strategic Games Mario Benevides Isaque Lima Rafael Nader Pedro Rougemont Systems and Computer Engineering Program and Computer Science Department Federal University of Rio de Janeiro, Brazil
More informationAn Improved Upper Bound for the Sum-free Subset Constant
1 2 3 47 6 23 11 Journal of Integer Sequences, Vol. 13 (2010), Article 10.8.3 An Improved Upper Bound for the Sum-free Subset Constant Mark Lewko Department of Mathematics University of Texas at Austin
More information