Different Modes of Discounting in Repeated Games

Total Page:16

File Type:pdf, Size:1020Kb

Different Modes of Discounting in Repeated Games Different Modes of Discounting in Repeated Games Stephen Wolff∗ under the guidance of J´er^omeRenault and Thomas Mariotti Toulouse School of Economics September 3, 2011 Abstract In this paper we study different modes of discounting in infinitely repeated games with perfect monitoring. In particular, we explore the equilibrium payoff set of the infinitely repeated symmetric two-player prisoner's dilemma under geometric and quasi-hyperbolic discounting. For each mode of discounting, we analyze the settings in which players have identical and different discount factors, and we provide an explicit characterization of a payoff profile on the Pareto frontier. 1 Introduction Repeated games have become a centerpiece in economic theory, describing repeated interac- tions among strategically thinking individuals. In most studies involving repeated games, two assumptions are made. The first is that players discount geometrically. Geometric discounting is one of the simplest and most mathematically rich modes of discounting. However, research suggests that humans discount hyperbolically (see Ainslie (1992)), assigning more weight to pay- offs that occur closer to the present, a phenomenon known as \present bias". Laibson (1997) popularized a family of discount functions that approximate hyperbolic discounting and that capture its key features while also allowing for tractable mathematical analysis. This family of functions, called \quasi-hyperbolic" discount functions, is the form usually encountered in the literature, and the form we will use in this paper. The second assumption commonly made in game-theoretic literature is that players use identical discount factors. While valid for certain settings, this assumption is by no means universally appropriate, and in fact it hides a time-dynamic aspect of the more general case. When players have different discount factors, the possibility of intertemporal trade becomes available. In particular, a relatively patient player can agree to grant a relatively impatient player higher rewards in earlier periods (which the impatient player weights more heavily) in exchange for higher rewards to the patient player in later periods (which the patient player weights more heavily). Such intertemporal trade can yield discounted total payoffs for the repeated game that lie outside the set of feasible rewards of the stage game. With hyperbolic discounting another novel aspect arises, namely, time inconsistency of players' planned actions. A player exhibiting present bias may plan to follow one strategy to maximize her total repeated-game payoff; but because of her present bias, when she actually finds herself in later periods, she may prefer to play a different strategy. This time inconsistency ∗Contact Stephen.Wolff@rice.edu with comments, corrections, or suggestions. 1 Discounting in Repeated Games Stephen Wolff, TSE (2011) LR T 1; 1 −1; 2 B 2; −1 0; 0 Figure 1: The Prisoner's Dilemma. means that subgame-perfect equilibria are no longer stable across time a priori, and we must refine our equilibrium concept, using instead the idea of Strotz-Pollak equilibria. This paper is organized as follows. We start by detailing the model we will use in our study. In section 3 we then analyze the model under geometric discounting, first laying out the well-studied baseline case of identical discount factors before moving to the case of different discount factors. Next, in section 4 we turn to hyperbolic discounting, again looking at the cases of identical and different discount factors. In section 5 we summarize our discussion and give some concluding remarks. 2 The Model Consider a stage game G = (N; (Ai)i2N ; (ui)i2N ), where N = f1; : : : ; ng is the (finite) set of players, Ai is the (finite) set of actions for player i, and ui : A ! R is the utility function for player i, describing the utility player i receives from each pure action profile in A = A1×· · ·×An. We allow players to use mixed actions in the stage game, so that the relevant stage game is the mixed extension of G, in which each player i chooses an action αi 2 ∆Ai and receives the expected stage-game utility, or reward, given by X ui(α) = ×j=1;:::;nαj(aj)ui(a); a2A where αj(aj) denotes the probability assigned by the mixed action αj to the pure action aj. The particular stage game that we will use throughout this study is the symmetric two-player prisoner's dilemma shown in Figure 1. We will assume the existence of a public randomization device; this ensures that the set of stage-game rewards (and later in the analysis, the set of continuation payoffs and the set of equilibrium payoffs) is convex. Let P denote the set of outcomes of the public randomization device. Given a public randomization device, the set V of feasible stage-game rewards is the convex hull of the set of pure-action rewards: V = convfu(a) j a 2 Ag: Let vi denote the minimax reward for player i in the stage game G: vi = min max ui(ai; α−i): α−i2∆A−i ai2Ai Let IRi = fv 2 V j vi ≥ vig denote the half-space of individually rational rewards for player i, and let IR denote the intersection of all such half-spaces for all players: \ IR = IRi = fv 2 V j vi ≥ vi8i 2 Ng: i2N 2 Discounting in Repeated Games Stephen Wolff, TSE (2011) ∗ Similarly, let IRi = fv 2 V j vi > vig denote the half-space of strictly individually rational rewards for player i, and let IR∗ denote the intersection of all such half-spaces for all players: ∗ \ ∗ IR = IRi = fv 2 V j vi > vi8i 2 Ng: i2N For the prisoner's dilemma in Figure 1, the set V of feasible stage-game rewards is the solid parallelogram with its four vertices at the four pure reward profiles. The minimax reward for each player i is vi = 0. The set IR of individually rational reward profiles is thus this paral- lelogram V intersected with the nonnegative quadrant, and the set IR∗ of strictly individually rational rewards is V intersected with the strictly positive quadrant. The stage game G is repeated infinitely many times, starting at period t = 0. We will assume perfect monitoring; that is, at the end of each stage, all players can perfectly observe the actions chosen in that stage (as well as the outcomes of all previous stages) and can condition their future actions on these outcomes. The stage-t history of the repeated game is a vector ht that includes all past actions of all players, i.e. (a0; : : : ; at−1) 2 At, and all past realized values of the public randomization device, i.e. (p0; : : : ; pt−1) 2 P t. We denote the set of all possible stage-t histories by Ht = At × P t. Players observe the realization of the public randomization device in period t before choosing t their actions, so that a stage-t pure strategy for player i is a map si : H × P ! Ai. 2.1 A Remark on Discounting In what follows we consider cases in which players use various modes of discounting to evaluate their payoff streams. However, we should note that discounting is not the only method for representing players' preferences over time. Osborne and Rubinstein (1994) detail two other methods. Under the limit of means criterion, player i strictly prefers the stream of rewards t +1 t +1 (vi )t=0 to the stream of rewards (wi)t=0 if and only if T X t t lim inf (vi − wi)=T > 0: T !+1 t=0 t +1 t +1 Under the overtaking criterion, player i strictly prefers the stream of rewards (vi )t=0 to (wi)t=0 if and only if T X t t lim inf (vi − wi) > 0: T !+1 t=0 Whereas discounting weights a given reward less and less the further away in time the reward is received (for discount factors less than one), the limit of means and overtaking criteria weight all time periods equally. There are convincing economic arguments for using discounting, and it is by far the most commonly encountered method of representing time-dependent preferences in the literature. Therefore, we will restrict our attention to discounting in the remainder of the paper. 3 Geometric Discounting Geometric discounting is the most widely used representation of time-dependent preferences. Under geometric discounting, player i has a discount factor δi 2 (0; 1); a reward obtained t 3 Discounting in Repeated Games Stephen Wolff, TSE (2011) t periods after the current period is multiplied by the weight δi . Note that, as t ! +1, the t weight δi ! 0, so that if the set Vi of player i's feasible stage-game rewards is bounded, then t +1 t the infinite sum of any reward stream (vi )t=0 is well-defined, where vi 2 Vi for all t. In practice, we usually define player i's payoff in the infinitely repeated game to be "+1 # X t t πi(σ) = (1 − δi)Eσ δi ui(a ) ; (1) t=0 where σ is any behavior-strategy profile of the infinitely repeated game and where the expec- tation is taken with respect to the probability distribution generated by σ over the (infinite) terminal histories. The normalization factor 1 − δi is included to render the payoffs of the infinitely repeated game directly comparable to the rewards of the stage game. Under geometric discounting, the repeated game assumes a symmetric structure through k time: The set of player i's continuation payoffs Ci in any future period k is identical to her set of feasible payoffs: " +1 # k X t−(k+1) t Ci (σ) = (1 − δi)Eσ δi ui(a ) : (2) t=k+1 The equivalence between the continuation payoff (2) and the feasible payoff (1) is easily seen by reindexing the sum in (2) with t0 = t − k.
Recommended publications
  • Lecture Notes
    GRADUATE GAME THEORY LECTURE NOTES BY OMER TAMUZ California Institute of Technology 2018 Acknowledgments These lecture notes are partially adapted from Osborne and Rubinstein [29], Maschler, Solan and Zamir [23], lecture notes by Federico Echenique, and slides by Daron Acemoglu and Asu Ozdaglar. I am indebted to Seo Young (Silvia) Kim and Zhuofang Li for their help in finding and correcting many errors. Any comments or suggestions are welcome. 2 Contents 1 Extensive form games with perfect information 7 1.1 Tic-Tac-Toe ........................................ 7 1.2 The Sweet Fifteen Game ................................ 7 1.3 Chess ............................................ 7 1.4 Definition of extensive form games with perfect information ........... 10 1.5 The ultimatum game .................................. 10 1.6 Equilibria ......................................... 11 1.7 The centipede game ................................... 11 1.8 Subgames and subgame perfect equilibria ...................... 13 1.9 The dollar auction .................................... 14 1.10 Backward induction, Kuhn’s Theorem and a proof of Zermelo’s Theorem ... 15 2 Strategic form games 17 2.1 Definition ......................................... 17 2.2 Nash equilibria ...................................... 17 2.3 Classical examples .................................... 17 2.4 Dominated strategies .................................. 22 2.5 Repeated elimination of dominated strategies ................... 22 2.6 Dominant strategies ..................................
    [Show full text]
  • Cooperative Strategies in Groups of Strangers: an Experiment
    KRANNERT SCHOOL OF MANAGEMENT Purdue University West Lafayette, Indiana Cooperative Strategies in Groups of Strangers: An Experiment By Gabriele Camera Marco Casari Maria Bigoni Paper No. 1237 Date: June, 2010 Institute for Research in the Behavioral, Economic, and Management Sciences Cooperative strategies in groups of strangers: an experiment Gabriele Camera, Marco Casari, and Maria Bigoni* 15 June 2010 * Camera: Purdue University; Casari: University of Bologna, Piazza Scaravilli 2, 40126 Bologna, Italy, Phone: +39- 051-209-8662, Fax: +39-051-209-8493, [email protected]; Bigoni: University of Bologna, Piazza Scaravilli 2, 40126 Bologna, Italy, Phone: +39-051-2098890, Fax: +39-051-0544522, [email protected]. We thank Jim Engle-Warnick, Tim Cason, Guillaume Fréchette, Hans Theo Normann, and seminar participants at the ESA in Innsbruck, IMEBE in Bilbao, University of Frankfurt, University of Bologna, University of Siena, and NYU for comments on earlier versions of the paper. This paper was completed while G. Camera was visiting the University of Siena as a Fulbright scholar. Financial support for the experiments was partially provided by Purdue’s CIBER and by the Einaudi Institute for Economics and Finance. Abstract We study cooperation in four-person economies of indefinite duration. Subjects interact anonymously playing a prisoner’s dilemma. We identify and characterize the strategies employed at the aggregate and at the individual level. We find that (i) grim trigger well describes aggregate play, but not individual play; (ii) individual behavior is persistently heterogeneous; (iii) coordination on cooperative strategies does not improve with experience; (iv) systematic defection does not crowd-out systematic cooperation.
    [Show full text]
  • Games with Delays. a Frankenstein Approach
    Games with Delays – A Frankenstein Approach Dietmar Berwanger and Marie van den Bogaard LSV, CNRS & ENS Cachan, France Abstract We investigate infinite games on finite graphs where the information flow is perturbed by non- deterministic signalling delays. It is known that such perturbations make synthesis problems virtually unsolvable, in the general case. On the classical model where signals are attached to states, tractable cases are rare and difficult to identify. Here, we propose a model where signals are detached from control states, and we identify a subclass on which equilibrium outcomes can be preserved, even if signals are delivered with a delay that is finitely bounded. To offset the perturbation, our solution procedure combines responses from a collection of virtual plays following an equilibrium strategy in the instant- signalling game to synthesise, in a Frankenstein manner, an equivalent equilibrium strategy for the delayed-signalling game. 1998 ACM Subject Classification F.1.2 Modes of Computation; D.4.7 Organization and Design Keywords and phrases infinite games on graphs, imperfect information, delayed monitoring, distributed synthesis 1 Introduction Appropriate behaviour of an interactive system component often depends on events gener- ated by other components. The ideal situation, in which perfect information is available across components, occurs rarely in practice – typically a component only receives signals more or less correlated with the actual events. Apart of imperfect signals generated by the system components, there are multiple other sources of uncertainty, due to actions of the system environment or to unreliable behaviour of the infrastructure connecting the compo- nents: For instance, communication channels may delay or lose signals, or deliver them in a different order than they were emitted.
    [Show full text]
  • Repeated Games
    Repeated games Felix Munoz-Garcia Strategy and Game Theory - Washington State University Repeated games are very usual in real life: 1 Treasury bill auctions (some of them are organized monthly, but some are even weekly), 2 Cournot competition is repeated over time by the same group of firms (firms simultaneously and independently decide how much to produce in every period). 3 OPEC cartel is also repeated over time. In addition, players’ interaction in a repeated game can help us rationalize cooperation... in settings where such cooperation could not be sustained should players interact only once. We will therefore show that, when the game is repeated, we can sustain: 1 Players’ cooperation in the Prisoner’s Dilemma game, 2 Firms’ collusion: 1 Setting high prices in the Bertrand game, or 2 Reducing individual production in the Cournot game. 3 But let’s start with a more "unusual" example in which cooperation also emerged: Trench warfare in World War I. Harrington, Ch. 13 −! Trench warfare in World War I Trench warfare in World War I Despite all the killing during that war, peace would occasionally flare up as the soldiers in opposing tenches would achieve a truce. Examples: The hour of 8:00-9:00am was regarded as consecrated to "private business," No shooting during meals, No firing artillery at the enemy’s supply lines. One account in Harrington: After some shooting a German soldier shouted out "We are very sorry about that; we hope no one was hurt. It is not our fault, it is that dammed Prussian artillery" But... how was that cooperation achieved? Trench warfare in World War I We can assume that each soldier values killing the enemy, but places a greater value on not getting killed.
    [Show full text]
  • Role Assignment for Game-Theoretic Cooperation ⋆
    Role Assignment for Game-Theoretic Cooperation ? Catherine Moon and Vincent Conitzer Duke University, Durham, NC 27708, USA Abstract. In multiagent systems, often agents need to be assigned to different roles. Multiple aspects should be taken into account for this, such as agents' skills and constraints posed by existing assignments. In this paper, we focus on another aspect: when the agents are self- interested, careful role assignment is necessary to make cooperative be- havior an equilibrium of the repeated game. We formalize this problem and provide an easy-to-check necessary and sufficient condition for a given role assignment to induce cooperation. However, we show that finding whether such a role assignment exists is in general NP-hard. Nevertheless, we give two algorithms for solving the problem. The first is based on a mixed-integer linear program formulation. The second is based on a dynamic program, and runs in pseudopolynomial time if the number of agents is constant. However, the first algorithm is much, much faster in our experiments. 1 Introduction Role assignment is an important problem in the design of multiagent systems. When multiple agents come together to execute a plan, there is generally a natural set of roles in the plan to which the agents need to be assigned. There are, of course, many aspects to take into account in such role assignment. It may be impossible to assign certain combinations of roles to the same agent, for example due to resource constraints. Some agents may be more skilled at a given role than others. In this paper, we assume agents are interchangeable and instead consider another aspect: if the agents are self-interested, then the assignment of roles has certain game-theoretic ramifications.
    [Show full text]
  • Interactively Solving Repeated Games. a Toolbox
    Interactively Solving Repeated Games: A Toolbox Version 0.2 Sebastian Kranz* March 27, 2012 University of Cologne Abstract This paper shows how to use my free software toolbox repgames for R to analyze infinitely repeated games. Based on the results of Gold- luecke & Kranz (2012), the toolbox allows to compute the set of pure strategy public perfect equilibrium payoffs and to find optimal equilib- rium strategies for all discount factors in repeated games with perfect or imperfect public monitoring and monetary transfers. This paper explores various examples, including variants of repeated public goods games and different variants of repeated oligopolies, like oligopolies with many firms, multi-market contact or imperfect monitoring of sales. The paper also explores repeated games with infrequent external auditing and games were players can secretly manipulate signals. On the one hand the examples shall illustrate how the toolbox can be used and how our algorithms work in practice. On the other hand, the exam- ples may in themselves provide some interesting economic and game theoretic insights. *[email protected]. I want to thank the Deutsche Forschungsgemeinschaft (DFG) through SFB-TR 15 for financial support. 1 CONTENTS 2 Contents 1 Overview 4 2 Installation of the software 5 2.1 Installing R . .5 2.2 Installing the package under Windows 32 bit . .6 2.3 Installing the package for a different operating system . .6 2.4 Running Examples . .6 2.5 Installing a text editor for R files . .7 2.6 About RStudio . .7 I Perfect Monitoring 8 3 A Simple Cournot Game 8 3.1 Initializing the game .
    [Show full text]
  • New Perspectives on Time Inconsistency
    Behavioural Macroeconomic Policy: New perspectives on time inconsistency Michelle Baddeley July 19, 2019 Abstract This paper brings together divergent approaches to time inconsistency from macroeconomic policy and behavioural economics. Developing Barro and Gordon’s models of rules, discretion and reputation, behavioural discount functions from behavioural microeconomics are embedded into Barro and Gordon’s game-theoretic analysis of temptation versus enforce- ment to construct an encompassing model, nesting combinations of time consistent and time inconsistent preferences. The analysis presented in this paper shows that, with hyperbolic/quasi-hyperbolic discounting, the enforceable range of inflation targets is narrowed. This suggests limits to the effectiveness of monetary targets, under certain conditions. The pa- per concludes with a discussion of monetary policy implications, explored specifically in the light of current macroeconomic policy debates. JEL codes: D03 E03 E52 E61 Keywords: time-inconsistency behavioural macroeconomics macroeconomic policy 1 Introduction Contrasting analyses of time inconsistency are presented in macroeconomic pol- icy and behavioural economics - the first from a rational expectations perspec- tive [12, 6, 7], and the second building on insights from behavioural economics arXiv:1907.07858v1 [econ.TH] 18 Jul 2019 about bounded rationality and its implications in terms of present bias and hyperbolic/quasi-hyperbolic discounting [13, 11, 8, 9, 18]. Given the divergent assumptions about rational choice across these two approaches, it is not so sur- prising that there have been few attempts to reconcile these different analyses of time inconsistency. This paper fills the gap by exploring the distinction between inter-personal time inconsistency (rational expectations models) and intra-personal time incon- sistency (behavioural economic models).
    [Show full text]
  • Cooperation in the Finitely Repeated Prisoner's Dilemma
    Cooperation in the Finitely Repeated Prisoner’s Dilemma Matthew Embrey Guillaume R. Fr´echette Sevgi Yuksel U. of Sussex ................................ NYU ................................... UCSB June, 2017 Abstract More than half a century after the first experiment on the finitely repeated prisoner’s dilemma, evidence on whether cooperation decreases with experience–as suggested by backward induction–remains inconclusive. This paper provides a meta-analysis of prior experimental research and reports the results of a new experiment to elucidate how coop- eration varies with the environment in this canonical game. We describe forces that a↵ect initial play (formation of cooperation) and unraveling (breakdown of cooperation). First, contrary to the backward induction prediction, the parameters of the repeated game have a significant e↵ect on initial cooperation. We identify how these parameters impact the value of cooperation–as captured by the size of the basin of attraction of Always Defect–to account for an important part of this e↵ect. Second, despite these initial di↵erences, the evolution of behavior is consistent with the unraveling logic of backward induction for all parameter combinations. Importantly, despite the seemingly contradictory results across studies, this paper establishes a systematic pattern of behavior: subjects converge to use threshold strategies that conditionally cooperate until a threshold round; and conditional on establishing cooperation, the first defection round moves earlier with experience. Sim- ulation results generated from a learning model estimated at the subject level provide insights into the long-term dynamics and the forces that slow down the unraveling of cooperation. We wish to thank the editor, anonymous referees, James Andreoni, Yoella Bereby-Meyer, Russell Cooper, Pedro Dal B´o, Douglas DeJong, Robert Forsythe, Drew Fudenberg, Daniel Friedman, P.J.
    [Show full text]
  • Lecture 19: Infinitely Repeated Games
    Lecture 19 Thursday, May 07, 2020 3:46 PM Lecture19… Lecture 19: Infinitely Repeated Games Mauricio Romero Lecture 19: Infinitely Repeated Games Introduction to Infinitely Repeated Games Lecture 19: Infinitely Repeated Games Introduction to Infinitely Repeated Games Ii'- One of the features of finitely repeated games was that if the stage game had a unique Nash equilibrium, then the only subgame perfect Nash equilibrium was the re petition of that unique stage game Nash equilibrium Iii- One of the features of finitely repeated games was that if the stage game had a unique Nash equilibrium, t hen t he only subgame perfect Nash equilibrium was the repetition of that unique stage game Nash equilibrium Iii- This happened because there was a last period from which we could induct backwards (and there was a domino effect!) Iii- One of the features of finitely repeated games was that if the stage game had a unique Nash equilibrium, t hen t he only subgame perfect Nash equilibrium was the repetition of that unique stage game Nash equilibrium Iii- This happened because there wa s a last period from whi ch we could induct backwards (and t here was a domino effect!) Iii- When the game is instead ~ted, this argument no longer applies since there is no such thing as a last period Iii- Lets first define what an infinitely repeated game is .,_ Lets first define what an infinitely repeated game is Iii- We start with a stage game whose utilities are given by u1 . u2 .. ... Un Iii- Lets first define what an infinitely repeated game is Iii- We start with a stage game whose utilities are given by u1.
    [Show full text]
  • Advanced Theory Field Exam, August 2013
    Theory Field Examination Behavioral economics (219A) Aug. 2013 Problem 1. Suppose that Jay (“J”)will live for T = 3 periods (periods 1, 2, and 3). In each of these three periods he chooses whether to consume sugary drinks, at = 1; or not to consume sugary drinks, at = 0: Sugary drinks are addictive, and J (who can get such drinks for free and has no other source of pleasure or displeasure in life) has preferences in each of periods 1, 2, and 3 of the form: ut(at = 0 at 1 = 0) = 0 j ut(at = 1 at 1 = 0) = 1 j ut(at = 0 at 1 = 1) = j 1 ut(at = 1 at 1 = 1) = 1 j These preferences capture the notion that consuming in the previous period makes J addicted: it means he gets less utility from continuing to consume sugary drinks than he does from consuming for the …rst time –he no longer gets the burst of pleasure and even gets a “hangover” – , but withdrawing is very unpleasant. J dies at the beginning of Period 4, and his utility from that point on doesn’tdepend on what he does in his lifetime. Importantly, assume that a0 = 0 in all parts below: J is born unaddicted. Suppose that J has (quasi-)hyperbolic discounting preferences with < 1 and = 1. J might be either sophisticated or naive. For all questions below, don’tworry about specifying behavior for any knife-edge values of that makes J indi¤erent among choices in any contingency. a) What will J do in Period 3, as a function of , whether he is sophisticated or naive, and whether he enters the period having chosen a2 = 0 or a2 = 1 in the previous period? Brie‡y give an intuition for you answer.
    [Show full text]
  • Evolutionary Game Theory: a Renaissance
    games Review Evolutionary Game Theory: A Renaissance Jonathan Newton ID Institute of Economic Research, Kyoto University, Kyoto 606-8501, Japan; [email protected] Received: 23 April 2018; Accepted: 15 May 2018; Published: 24 May 2018 Abstract: Economic agents are not always rational or farsighted and can make decisions according to simple behavioral rules that vary according to situation and can be studied using the tools of evolutionary game theory. Furthermore, such behavioral rules are themselves subject to evolutionary forces. Paying particular attention to the work of young researchers, this essay surveys the progress made over the last decade towards understanding these phenomena, and discusses open research topics of importance to economics and the broader social sciences. Keywords: evolution; game theory; dynamics; agency; assortativity; culture; distributed systems JEL Classification: C73 Table of contents 1 Introduction p. 2 Shadow of Nash Equilibrium Renaissance Structure of the Survey · · 2 Agency—Who Makes Decisions? p. 3 Methodology Implications Evolution of Collective Agency Links between Individual & · · · Collective Agency 3 Assortativity—With Whom Does Interaction Occur? p. 10 Assortativity & Preferences Evolution of Assortativity Generalized Matching Conditional · · · Dissociation Network Formation · 4 Evolution of Behavior p. 17 Traits Conventions—Culture in Society Culture in Individuals Culture in Individuals · · · & Society 5 Economic Applications p. 29 Macroeconomics, Market Selection & Finance Industrial Organization · 6 The Evolutionary Nash Program p. 30 Recontracting & Nash Demand Games TU Matching NTU Matching Bargaining Solutions & · · · Coordination Games 7 Behavioral Dynamics p. 36 Reinforcement Learning Imitation Best Experienced Payoff Dynamics Best & Better Response · · · Continuous Strategy Sets Completely Uncoupled · · 8 General Methodology p. 44 Perturbed Dynamics Further Stability Results Further Convergence Results Distributed · · · control Software and Simulations · 9 Empirics p.
    [Show full text]
  • An Introduction to Game Theory
    Publicly•available solutions for AN INTRODUCTION TO GAME THEORY Publicly•available solutions for AN INTRODUCTION TO GAME THEORY MARTIN J. OSBORNE University of Toronto Copyright © 2012 by Martin J. Osborne All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of Martin J. Osborne. This manual was typeset by the author, who is greatly indebted to Donald Knuth (TEX), Leslie Lamport (LATEX), Diego Puga (mathpazo), Christian Schenk (MiKTEX), Ed Sznyter (ppctr), Timothy van Zandt (PSTricks), and others, for generously making superlative software freely available. The main font is 10pt Palatino. Version 6: 2012-4-7 Contents Preface xi 1 Introduction 1 Exercise 5.3 (Altruistic preferences) 1 Exercise 6.1 (Alternative representations of preferences) 1 2 Nash Equilibrium 3 Exercise 16.1 (Working on a joint project) 3 Exercise 17.1 (Games equivalent to the Prisoner’s Dilemma) 3 Exercise 20.1 (Games without conflict) 3 Exercise 31.1 (Extension of the Stag Hunt) 4 Exercise 34.1 (Guessing two-thirds of the average) 4 Exercise 34.3 (Choosing a route) 5 Exercise 37.1 (Finding Nash equilibria using best response functions) 6 Exercise 38.1 (Constructing best response functions) 6 Exercise 38.2 (Dividing money) 7 Exercise 41.1 (Strict and nonstrict Nash equilibria) 7 Exercise 47.1 (Strict equilibria and dominated actions) 8 Exercise 47.2 (Nash equilibrium and weakly dominated
    [Show full text]