Combinatorial Algorithms for Convex Programs Algorithmic Game Theory (Capturing Market Equilibria and and Internet Computing Nash Bargaining Solutions) Vijay V. Vazirani Georgia Tech What is Economics? ‘‘Economics is the study of the use of scarce resources which have alternative uses.’’ Lionel Robbins (1898 – 1984) How are scarce resources assigned to alternative uses? How are scarce resources assigned to alternative uses? Prices! How are scarce resources assigned to alternative uses? Prices Parity between demand and supply How are scarce resources assigned to alternative uses? Prices Parity between demand and supply equilibrium prices Do markets even admit equilibrium prices? Do markets even admit equilibrium prices? General Equilibrium Theory Occupied center stage in Mathematical Economics for over a century Arrow-Debreu Theorem, 1954 Celebrated theorem in Mathematical Economics Established existence of market equilibrium under very general conditions using a theorem from topology - Kakutani fixed point theorem. Do markets even admit equilibrium prices? Do markets even admit equilibrium prices? Easy if only one good! Supply-demand curves Do markets even admit equilibrium prices? What if there are multiple goods and multiple buyers with diverse desires and different buying power? Irving Fisher, 1891 Defined a fundamental market model linear utilities vi uij xij jG xij 0 For given prices, find optimal bundle of goods p1 p2 p3 Several buyers with different utility functions and moneys. Several buyers with different utility functions and moneys. Find equilibrium prices. p1 p2 p3 Arrow-Debreu Theorem, 1954 Celebrated theorem in Mathematical Economics Established existence of market equilibrium under very general conditions using a theorem from topology - Kakutani fixed point theorem. Highly non-constructive! General Equilibrium Theory An almost entirely non-algorithmic theory! The new face of computing Today’s reality New markets defined by Internet companies, e.g., Google eBay Yahoo! Amazon Massive computing power available. Need an inherenltly-algorithmic theory of markets and market equilibria. Combinatorial Algorithm for Linear Case of Fisher’s Model Devanur, Papadimitriou, Saberi & V., 2002 Using the primal-dual paradigm Combinatorial algorithm Conducts an efficient search over a discrete space. E.g., for LP: simplex algorithm vs ellipsoid algorithm or interior point algorithms. Combinatorial algorithm Conducts an efficient search over a discrete space. E.g., for LP: simplex algorithm vs ellipsoid algorithm or interior point algorithms. Yields deep insights into structure. No LP’s known for capturing equilibrium allocations for Fisher’s model No LP’s known for capturing equilibrium allocations for Fisher’s model Eisenberg-Gale convex program, 1959 Eisenberg-Gale Program, 1959 max mi log vi i s.t. i : vi j u ij x ij x 1 ij : x 0 j : ij i ij Eisenberg-Gale Program, 1959 max mi log vi i s.t. i : vi j u ij x ij x 1 ij : x 0 j : ij i ij prices pj No LP’s known for capturing equilibrium allocations for Fisher’s model Eisenberg-Gale convex program, 1959 Extended primal-dual paradigm to solving a nonlinear convex program Theorem If all parameters are rational, Eisenberg-Gale convex program has a rational solution! Polynomially many bits in size of instance Theorem If all parameters are rational, Eisenberg-Gale convex program has a rational solution! Polynomially many bits in size of instance Combinatorial polynomial time algorithm for finding it. Theorem If all parameters are rational, Eisenberg-Gale convex program has a rational solution! Polynomially many bits in size of instance Combinatorial polynomial time algorithm for finding it. Discrete space Idea of algorithm primal variables: allocations dual variables: prices of goods iterations: execute primal & dual improvements Allocations Prices (Money) How are scarce resources assigned to alternative uses? Prices Parity between demand and supply Yin & Yang Nash bargaining game, 1950 Captures the main idea that both players gain if they agree on a solution. Else, they go back to status quo. Complete information game. Example Two players, 1 and 2, have vacation homes: 1: in the mountains 2: on the beach Consider all possible ways of sharing. Utilities derived jointly v2 S : convex + compact feasible set v1 Disagreement point = status quo utilities v2 S c2 c1 Disagreement point = (c1 , c2 ) v1 Nash bargaining problem = (S, c) v2 S c2 c1 Disagreement point = (c1 , c2 ) v1 Nash bargaining Q: Which solution is the “right” one? Solution must satisfy 4 axioms: Paretto optimality Invariance under affine transforms Symmetry Independence of irrelevant alternatives Thm: Unique solution satisfying 4 axioms N ( S , c) max ( v1 ,v2 )S {(v1 c1 )(v2 c2 )} v2 S c2 c1 v1 Generalizes to n-players Theorem: Unique solution N (S , c) max vS {(v1 c1 ) ... (vn cn )} Linear Nash Bargaining (LNB) Feasible set is a polytope defined by linear packing constraints Nash bargaining solution is optimal solution to convex program: max log(vi ci ) i s.t. packing constraints Q: Compute solution combinatorially in polynomial time? How should they exchange their goods? State as a Nash bargaining game u f : (.,.,.) R ub : (.,.,.) R um : (.,.,.) R c f u f (1, 0, 0) cb ub (0, 1, 0) cm um (0, 0,1) S = utility vectors obtained by distributing goods among players Special case: linear utility functions u f : (.,.,.) R ub : (.,.,.) R um : (.,.,.) R c f u f (1, 0, 0) cb ub (0, 1, 0) cm um (0, 0,1) S = utility vectors obtained by distributing goods among players Convex program for ADNB max log(vi ci ) i s.t. i : vi j u ij x ij j : ij : x 1 x 0 i ij ij Theorem (V., 2008) If all parameters are rational, solution to ADNB is rational! Polynomially many bits in size of instance Theorem (V., 2008) If all parameters are rational, solution to ADNB is rational! Polynomially many bits in size of instance Combinatorial polynomial time algorithm for finding it. Flexible budget markets Natural variant of linear Fisher markets ADNB flexible budget markets Primal-dual algorithm for finding an equilibrium How is primal-dual paradigm adapted to nonlinear setting? Fundamental difference between LP’s and convex programs Complementary slackness conditions: involve primal or dual variables, not both. KKT conditions: involve primal and dual variables simultaneously. KKT conditions 1.j : p j 0 2.j : p j 0 i xij 1 uij vi 3.i, j : p j m(i ) uij vi 4.i, j : xij 0 p j m(i ) KKT conditions 1.j : p j 0 2.j : p j 0 i xij 1 uij vi 3.i, j : p j m(i ) uij vi 4.i, j : xij 0 p j m(i ) u x ij ij j m(i ) Primal-dual algorithms so far (i.e., LP-based) Raise dual variables greedily. (Lot of effort spent on designing more sophisticated dual processes.) Primal-dual algorithms so far Raise dual variables greedily. (Lot of effort spent on designing more sophisticated dual processes.) Only exception: Edmonds, 1965: algorithm for max weight matching. Primal-dual algorithms so far Raise dual variables greedily. (Lot of effort spent on designing more sophisticated dual processes.) Only exception: Edmonds, 1965: algorithm for max weight matching. Otherwise primal objects go tight and loose. Difficult to account for these reversals -in the running time. Our algorithm Dual variables (prices) are raised greedily Yet, primal objects go tight and loose Because of enhanced KKT conditions Our algorithm Dual variables (prices) are raised greedily Yet, primal objects go tight and loose Because of enhanced KKT conditions New algorithmic ideas needed! Open Nonlinear programs with rational solutions! Open Nonlinear programs with rational solutions! Solvable combinatorially!! Primal-Dual Paradigm Combinatorial Optimization (1960’s & 70’s): Integral optimal solutions to LP’s Exact Algorithms for Cornerstone Problems in P Matching (general graph) Network flow Shortest paths Minimum spanning tree Minimum branching Primal-Dual Paradigm Combinatorial Optimization (1960’s & 70’s): Integral optimal solutions to LP’s Approximation Algorithms (1990’s): Near-optimal integral solutions to LP’s Approximation Algorithms set cover Steiner tree Steiner network k-MST scheduling . . . facility location k-median multicut feedback vertex set Primal-Dual Paradigm Combinatorial Optimization (1960’s & 70’s): Integral optimal solutions to LP’s Approximation Algorithms (1990’s): Near-optimal integral solutions to LP’s Algorithmic Game Theory (New Millennium): Rational solutions to nonlinear convex programs Primal-Dual Paradigm Combinatorial Optimization (1960’s & 70’s): Integral optimal solutions to LP’s Approximation Algorithms (1990’s): Near-optimal integral solutions to LP’s Algorithmic Game Theory (New Millennium): Rational solutions to nonlinear convex programs Approximation algorithms for convex programs?! Convex program for ADNB max log(vi ci ) i s.t. i : vi j u ij x ij j : ij : x 1 x 0 i ij ij Eisenberg-Gale Program, 1959 max mi log vi i s.t. i : vi j u ij x ij x 1 ij : x 0 j : ij i ij Common generalization max wi log(vi ci ) i s.t. i : vi j u ij x ij j : ij : x 1 x 0 i ij ij Common generalization Is it meaningful? Can it be solved via a combinatorial, polynomial time algorithm? Common generalization Is it meaningful? Kalai, 1975: Nonsymmetric bargaining games wi : clout of player i. Nonsymmetric ADNB Common generalization Is it meaningful? Kalai, 1975: Nonsymmetric bargaining games wi : clout of player i. Algorithm Nonsymmetric ADNB