Top Banner
23

Approximations and Truthfulness: The Case of Multi-Unit Auctions

Dec 31, 2015

Download

Documents

blossom-price

Approximations and Truthfulness: The Case of Multi-Unit Auctions. Shahar Dobzinski Joint work with Noam Nisan and with Shaddin Dughmi. Auctions. Clean Air Auction. 1990’s: The US government decided to decrease the atmospheric levels of sulfur dioxide. - PowerPoint PPT Presentation
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Approximations and Truthfulness: The Case of Multi-Unit Auctions
Page 2: Approximations and Truthfulness: The Case of Multi-Unit Auctions

Auctions

Page 3: Approximations and Truthfulness: The Case of Multi-Unit Auctions

Clean Air Auction

Page 4: Approximations and Truthfulness: The Case of Multi-Unit Auctions

Definition of Multi-Unit Auctions

Page 5: Approximations and Truthfulness: The Case of Multi-Unit Auctions

Can we obtain an algorithm that is fast, truthful and approximates the welfare

well?

Page 6: Approximations and Truthfulness: The Case of Multi-Unit Auctions

Related Work

Page 7: Approximations and Truthfulness: The Case of Multi-Unit Auctions

Theorem: There exists a truthful poly time randomized (1-)-approximation mechanism.

Theorem: There exists a poly time deterministic truthful ½-approximation mechanism.

Page 8: Approximations and Truthfulness: The Case of Multi-Unit Auctions

Maximal in Range: limit the range and fully optimize over the restricted range.

Page 9: Approximations and Truthfulness: The Case of Multi-Unit Auctions

The mechanism: (1)Split the items into n2 equi-sized bundles each of size m/n2.(2)Allocate these bundles optimally.

Page 10: Approximations and Truthfulness: The Case of Multi-Unit Auctions
Page 11: Approximations and Truthfulness: The Case of Multi-Unit Auctions

Theorem: Every maximal-in-range (½+)-approx algorithm requires at least m queries to the black boxes.

Page 12: Approximations and Truthfulness: The Case of Multi-Unit Auctions
Page 13: Approximations and Truthfulness: The Case of Multi-Unit Auctions

Proving Impossibilities:Characterize and Optimize

+ =

Page 14: Approximations and Truthfulness: The Case of Multi-Unit Auctions

What’s Next?

Theorem: there exists a (1-)-approximation mechanism that is truthful in expectation.

Page 15: Approximations and Truthfulness: The Case of Multi-Unit Auctions
Page 16: Approximations and Truthfulness: The Case of Multi-Unit Auctions
Page 17: Approximations and Truthfulness: The Case of Multi-Unit Auctions

The AlgorithmSome valuation function

Items

Value

Page 18: Approximations and Truthfulness: The Case of Multi-Unit Auctions

Correctness

Page 19: Approximations and Truthfulness: The Case of Multi-Unit Auctions

Summary

Page 20: Approximations and Truthfulness: The Case of Multi-Unit Auctions

Open Questions

Page 21: Approximations and Truthfulness: The Case of Multi-Unit Auctions

Basic Requirements from Auctions

This talk: auction design via multi-unit auctions.

Page 22: Approximations and Truthfulness: The Case of Multi-Unit Auctions

Modern Auctions I: eBay

Page 23: Approximations and Truthfulness: The Case of Multi-Unit Auctions

Modern Auctions II: Sponsored Search