Presentation is loading. Please wait.

Presentation is loading. Please wait.

Comp 553: Algorithmic Game Theory Fall 2014 Yang Cai Lecture 23.

Similar presentations


Presentation on theme: "Comp 553: Algorithmic Game Theory Fall 2014 Yang Cai Lecture 23."— Presentation transcript:

1 Comp 553: Algorithmic Game Theory Fall 2014 Yang Cai Lecture 23

2 Exchange Market Model traders divisible goods trader i has: - endowment of goods non-negative reals amount of goods trader comes to the marketplace with consumption set for trader i specifies trader i’s utility for bundles of goods - utility function

3 Fisher Market Model can be obtained as a special case of an exchange market, when endowment vectors are parallel: in this case, relative incomes of the traders are independent of the prices. n traders with: k divisible goods owned by seller; money m i, and utility function u i seller has q j units of good j

4 Competitive (or Walrasian) Market Equilibrium total demand total supply Def: A price vector is called a competitive market equilibrium iff there exists a collection of optimal bundles of goods, for all traders i = 1,…, n, such that the total supply meets the total demand, i.e. [ For Fisher Markets: ]

5 Arrow-Debreu Theorem 1954 Theorem [Arrow-Debreu 1954]: Suppose Then a competitive market equilibrium exists. (i) is closed and convex (iii a) (iii b) (iii c) (ii) (all coordinates positive)

6 Fisher’s Model Suppose all endowment vectors are parallel… Equivalently, we can imagine the following situation:  relative incomes of the traders are independent of the prices. n traders, with specified money m i Arrow-Debreu Thm  (under the Arrow-Debreu conditions) there exist prices that the seller can assign on the goods so that the traders spend all their money to buy optimal bundles and supply meets demand k divisible goods owned by seller; seller has q j units of good j

7 Computation

8 Utility Functions CES (Constant Elasticity of Substitution) utility functions: linear utility form Leontief utility form Cobb-Douglas form

9  Buyers’ optimization program (under price vector p): Fisher’s Model with CES utility functions  Global Constraint:

10  The space of feasible allocations is: Eisenberg-Gale’s Convex Program e.g., choosing as a global objective function the sum of the traders’ utility functions won’t work…  But how do we aggregate the trader’s optimization problems into one global optimization problem?

11 Observation: The global optimization problem should not favor (or punish) Buyer i should he  Eisenberg and Gale’s idea: Use the following objective function (take its logarithm to convert into a concave function) Eisenberg-Gale’s Convex Program  Split himself into two buyers with half the money  Double all her u ij ’ s

12 Eisenberg-Gale’s Convex Program Remarks:- No budget constraints! - It is not necessary that the utility functions are CES; everything works as long as they are concave, and homogeneous

13 Eisenberg-Gale’s Convex Program KKT Conditions  - primal variables + Langrange multipliers comprise a competitive eq. - interpret Langrange multipliers as prices 1. Gives a poly-time algorithm for computing a market equilibrium in Fisher’s model. 2. At the same time provides a proof that a market equilibrium exists in this model. Linear case: proof on the board.


Download ppt "Comp 553: Algorithmic Game Theory Fall 2014 Yang Cai Lecture 23."

Similar presentations


Ads by Google