Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked...
-
Upload
travis-trew -
Category
Documents
-
view
218 -
download
3
Transcript of Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked...
![Page 1: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/1.jpg)
Algorithmic Game Theoryand Internet Computing
Amin Saberi
Algorithmic Game Theory and Networked Systems
![Page 2: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/2.jpg)
Outline
Game Theory and Algorithmsefficient algorithms for game theoretic notions
Complex networks andperformance of basic algorithms
![Page 3: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/3.jpg)
Internet
Algorithms Game Theory efficiency,
distributed computation, scalability
Huge, distributed, dynamicowned and operated by different entities
strategies, fairness, incentive compatibility
![Page 4: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/4.jpg)
Internet
Algorithmic Game Theory efficiency,
distributed computation, scalability
strategies, fairness, incentive compatibility
Fusing ideas of algorithms and game theory
Huge, distributed, dynamicowned and operated by different entities
![Page 5: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/5.jpg)
Find efficient algorithms for computing game theoretic notions like:
Market equilibria Nash equilibria
Core Shapley value
Auction ...
![Page 6: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/6.jpg)
Find efficient algorithms for computing game theoretic notions like:
Market equilibria Nash equilibria
Core Shapley value
Auction ...
![Page 7: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/7.jpg)
Market Equilibrium Arrow-Debreu (1954): Existence of equilibrium prices
(highly non-constructive, using Kakutani’s fixed-point theorem)
Fisher (1891): Hydraulic apparatus for calculating equilibrium Eisenberg & Gale (1959): Convex program (ellipsoid implicit) Scarf (1973): Approximate fixed point algorithms
Use techniques from modern theory of algorithmsIs linear case in P? Deng, Papadimitriou, Safra (2002)
![Page 8: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/8.jpg)
![Page 9: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/9.jpg)
n buyers, with specified money m divisible goods (unit amount) Linear utilities: uij utility derived by i on obtaining one unit of j
Find prices such that buyers spend all their money Market clears
Market Equilibrium
![Page 10: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/10.jpg)
Buyer i’s optimization program:
Global Constraint:
Market Equilibrium
![Page 11: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/11.jpg)
Market Equilibrium
![Page 12: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/12.jpg)
Devanur, Papadimitriou, S., Vazirani ‘03: A combinatorial (primal-dual) algorithm for finding the prices in polynomial time.
![Page 13: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/13.jpg)
Devanur, Papadimitriou, S., Vazirani ‘03: A combinatorial (primal-dual) algorithm for finding the prices in polynomial time.
Start with small prices so that only buyers have surplus gradually increase prices until surplus is zero
Primal-dual scheme:
Allocation Prices
![Page 14: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/14.jpg)
Devanur, Papadimitriou, S., Vazirani ‘03: A combinatorial (primal-dual) algorithm for finding the prices in polynomial time.
Start with small prices so that only buyers have surplus gradually increase prices until surplus is zero
Primal-dual scheme:
Allocation Prices
Measure of progress: l2-norm of the surplus vector.
![Page 15: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/15.jpg)
Buyers Goods
$20 10/20
$40 20/40
$10 4/10
$60 2/60
10
20
4
2
Bang per buck: utility of worth $1 of a good.
Equality Subgraph
![Page 16: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/16.jpg)
Equality Subgraph Buyers Goods
Bang per buck: utility of worth $1 of a good.
Buyers will only buy the goods with highest bang per buck:
$20 10/20
$40 20/40
$10 4/10
$60 2/60
10
20
4
2
![Page 17: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/17.jpg)
Buyers Goods
Bang per buck: utility of worth $1 of a good.
Buyers will only buy the goods with highest bang per buck:
Equality Subgraph
![Page 18: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/18.jpg)
Buyers Goods
$20
$40
$10
$60
$100
$60
$20
$140
Buyers Goods
How do we compute the sales in equality subgraph ?
Equality Subgraph
![Page 19: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/19.jpg)
Buyers Goods
How do we compute the sales in equality subgraph ?
maximum flow!
20
40
10
60
100
60
20
140
Equality Subgraph
![Page 20: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/20.jpg)
Buyers Goods
100
60
20
140
always saturated (by invariant)
How do we compute the sales in equality subgraph ?
maximum flow!
Equality Subgraph
![Page 21: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/21.jpg)
Buyers Goods
buyers may have always saturated surplus (by invariant)
How do we compute the sales in equality subgraph ?
maximum flow!
Equality Subgraph
![Page 22: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/22.jpg)
100
60
20
140
20
40
10
60
Example
![Page 23: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/23.jpg)
20/100
20/60
20/20
70/140
Example
20/20
40/40
10/10
60/60
10
60
20
20
20
![Page 24: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/24.jpg)
20/100
20/60
20/20
70/140
Example
20/20
40/40
10/10
60/60
10
60
20
20
20
Surplus vector = (80, 40, 0, 70)
![Page 25: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/25.jpg)
Balanced FlowBalanced Flow: the flow that minimizes the l2-norm of the
surplus vector.
tries to make surplus of buyers as equal as possible
Theorem: Flow f is balanced iff there is no path from i to j with surplus(i) < surplus(j) in the residual graph corresponding to f .
![Page 26: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/26.jpg)
20/100
20/60
20/20
70/140
Example
20/20
40/40
10/10
60/60
10
60
20
20
20
Surplus vector = (80, 40, 0, 70)
![Page 27: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/27.jpg)
50/100
10/60
0/20
70/140
Example
20/20
40/40
10/10
60/60
10
60
30
10
20
(80, 40, 0, 70) (50, 50, 20, 70)
70/140
![Page 28: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/28.jpg)
How to raise the prices? Raise prices proportionately
Which goods ? goods connected to the buyers with
the highest surplus
i
j
l
j ij
l il
p u
p u
![Page 29: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/29.jpg)
l2-norm of the surplus vector decreases:
total surplus decreases
Flow becomes more balanced
Number of max-flow computations:
Buyers Goods
2 2( ( log log ))O n n U Mn
Algorithm
![Page 30: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/30.jpg)
Buyers Goods
l2-norm of the surplus vector decreases:
total surplus decreases
Flow becomes more balanced
Number of max-flow computations:
2 2( ( log log ))O n n U Mn
Algorithm
![Page 31: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/31.jpg)
Jain, Mahdian, S. (‘03): people can sell and buy at the same time
Kakade, Kearns, Ortiz (‘03): Graphical economics
Kapur, Garg (‘04): Auction Algorithm
Devanur, Vazirani (‘04): Spending Constraint
Jain (‘04), Ye (‘04): Non-linear program for a general case
Chen, Deng, Sun, Yao (‘04): Concave utility functions
Further work, extensions
![Page 32: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/32.jpg)
Primal-dual scheme (Kelly, Low, Tan, …): primal: packet rates at sources dual: congestion measures (shadow prices)
A market equilibrium in a distributed setting!
Congestion Control
![Page 33: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/33.jpg)
Next step: dealing with integral goods
Approximate equilibria:
Surplus or deficiency unavoidable
Minimize the surplus: NP-hard (approximation algorithm)
Fair allocation:
Max-min fair allocation (maximize minimum happiness)
Minimize the envy (Lipton, Markakis, Mossel, S. ‘04)
![Page 34: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/34.jpg)
Core of a Game
•
•
•
••
V(S): the total gain of playing the game within subset S
Core: Distribute the gain such that no subset has an incentive to secede:
• ••
N
S
![Page 35: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/35.jpg)
Stability in Routing•
•
•
•
••
Node i has capacity Ci
Demand dij between nodes i and j
Total gain of subset S, V(S):maximum amount of flow you can route in the graph induced by S. (solution of multi-commodity flow LP)
S
![Page 36: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/36.jpg)
Stability in Routing
Markakis, S. (‘03): The core of this game is nonempty i.e. there is a way to distribute the gain s.t.:
For all S:
Proof idea: Use the dual of the multi-commodity flow LP.
•
•
•
•
••
•
•
•
•
••
S
![Page 37: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/37.jpg)
Optimal Auction Design
Design an auction:
truthful
maximizes the expected revenue
AuctionProbability distribution
Buyer 1
Buyer 2
.
.
Buyer n
BidsQueries from oracleOracle
![Page 38: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/38.jpg)
Optimal Auction Design
History: Independent utilities: characterization
(Myerson, 1981) General case: factor ½-approximation
(Ronen , ‘01)
Ronen & S. (‘02): Hardness of approximation
Idea: Probabilistic construction; polynomial number of queries does not provide enough information!
![Page 39: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/39.jpg)
Computer Science Applications: networking: algorithms and protocols E–commerce
Game Theory Applications: modeling, simulation intrinsic complexity bounded rationality
![Page 40: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/40.jpg)
Outline
Game Theory and Algorithmsefficient algorithms for game theoretic notions
Complex networks and performance of basic algorithms
![Page 41: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/41.jpg)
Power Law Random Graphs Bollobas 80’s, Molloy&Reed 90’s,
Chung 00’s.
Preferential Attachment Simon 55, Barabasi-Albert 99, Kumar et al 00, Bollobas-Riordan 00, Bollobas et al 03.
The Internet Graph
Power Law:
![Page 42: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/42.jpg)
Power Law Random Graphs Bollobas 80’s, Molloy&Reed 90’s,
Chung 00’s.
Preferential Attachment Simon 55, Barabasi-Albert 99, Kumar et al 00, Bollobas-Riordan 00, Bollobas et al 03.
The Internet Graph
Power Law:
What is the impact of the structure on the performance of the system ?
![Page 43: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/43.jpg)
improving Cooper-Frieze for d large
In both models conductance a.s.
Power-law Random Graphs: when min degree ≥ 3.(Gkantisidis, Mihail, S. ‘03)
Preferential attachment: when d ≥ 2(Mihail, Papadimitriou, S. ‘03)
Main Implication: a.s.
Conductance and Eigenvalue gap
![Page 44: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/44.jpg)
# Cut edges
Conductance
![Page 45: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/45.jpg)
Routing with low congestion: We can route di dj flow between all vertices i and j
with maximum congestion at most O(n log n). (by approx. multicommodity flow Leighton-Rao ‘88)
Algorithmic Applications
![Page 46: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/46.jpg)
Bounds on mixing time, hitting time and cover time (searching and crawling)
Random walks for search and construction in P2P networks (Gkantsidis, Mihail, S. ‘04)
Search with replication in P2P networks generalizing the notion of cover time (Mihail, Tetali, S., in preparation)
Algorithmic Applications
![Page 47: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/47.jpg)
Absence of epidemic threshold in scale-free graphs (Berger, Borgs, Chayes, S., ‘04)
infected healthy at rate 1
healthy infected at rate ( )
Epidemic threshold :
In scale-free graphs, this threshold is zero!
Recent work
#infectedneighbors
![Page 48: Algorithmic Game Theory and Internet Computing Amin Saberi Algorithmic Game Theory and Networked Systems.](https://reader036.fdocuments.in/reader036/viewer/2022062511/551a3c1d550346a4248b590f/html5/thumbnails/48.jpg)
THE END !