Cognitive Radio Communications and Networks: Principles and Practice By A. M. Wyglinski, M. Nekovee, Y. T. Hou (Elsevier, December 2009) 1 Chapter 17 Auction-based.

“Cognitive Radio Communications and Networks: Principles and Practice”By A. M. Wyglinski, M. Nekovee, Y. T. Hou (Elsevier, December 2009)

1

Chapter 17

Auction-based spectrum markets in cognitive radio networks


2

Outline Rethinking Spectrum Auctions On-demand Spectrum Auctions Economic-Robust Spectrum Auctions Double Spectrum Auctions for Multi-party Trading Chapter Summary Further Reading


Recent Spectrum Auction Activities

1. Allocate spectrum statically in long-term (10 years) national leases2. Take months/years to complete

3. Expensive4. Controlled by incumbents (Verizon, AT&T)


Addressing Inefficient Spectrum Distribution

Legacy wireless providers own the majority of spectrum But cannot fully utilize it

New wireless providers are dying for usable spectrum But have to crowd into

limited unlicensed bands

Market-based Spectrum Trading

Market-based Spectrum Trading

SellersSellers

BuyersBuyers


Rethinking Spectrum Auctions

eBay in the Sky On-demand spectrum auctions

Short-term, local area, low-cost No need to pay for 10 years of

spectrum usage across the entire west-coast

Support small players and new market entrants

Stimulate fast innovations

Dynamic Spectrum Auctions

1

6

2

3

5 4


Why Auctions?

• Auctioneers periodically auction spectrum based on user bids Dynamically discover prices

based on demands

• Users request spectrum when they need it Match traffic dynamics Flexible and cost-effective

Dynamic Spectrum Auctions

1

6

2

3

5 4


Summary of Challenges

Multi-unit auctions Multiple winners Each assigned with a portion of

spectrum

Subject to interference constraints Combinatorial constraints among

bidders Complexity grows exponentially with

the number of bidders

NP-hard resource allocation problem

NP-hard resource allocation problem

Can we design low-complexity and yet efficient auction solutions for large scale systems?

Large # of

bidders

Real-time auctions


8



System Overview

Piecewise Linear Price Demand bids– a compact and yet highly expressive

bidding format

User Auctioneer

Uniform vs. Discriminatory pricing models – tradeoffs

between efficiency and fairness

BiddingBidding Pricing ModelPricing Model

Fast auction clearing algorithms for both pricing

models

Allocation (clearing)Allocation (clearing)

5

1

6 23 4

How do users bid?

How to set prices?

how to handle the bids to efficiently maximize

revenue?


Fast Auction Clearing

The problem is NP-hard because: Pair-wise combinatorial interference constraints

What if: convert the interference constraints into a set of linear constraints Functions of Xi: The amount of spectrum

assigned to bidder i Must be as strict as before Reduce the problem into variants of Linear

Programming Problem Can do this in a central controller

We propose: Node-L constraints

Original interference constraints


Analytical Bounds

CAUP Clearing Algorithm for Uniform Pricing

UPOPTCAUP RR 3

1

)loglog( UnnnO

CADP Clearing Algorithm for Discriminatory Pricing

DPOPTCADP Rn

nR

)( 13

polynomial

Revenue efficiency

Complexity

When the conflict graph

is a treeUPOPTCAUP RR DPOPTCADP RR

Theoretical bounds


As a Result…..

Using a normal desktop computer:

• An auction with 4000 bidders takes 90 seconds 20,000 time faster than the optimal solution

• If <100 bidders, only 15% revenue degradation over the optimal solution

Using a normal desktop computer:

• An auction with 4000 bidders takes 90 seconds 20,000 time faster than the optimal solution

• If <100 bidders, only 15% revenue degradation over the optimal solution


13




VERITAS: Truthful and Efficient Spectrum Auctions

VERITAS-Allocation: Bid-dependent greedy allocation Best known polynomial-time channel allocation schemes are greedy Enable spatial reuse Within a provable distance (Δ: max conflict degree) to the optimal

auction efficiency VERITAS-Pricing:

Charge every winner i, the bid of its critical neighbor C(i) Critical Neighbor: The neighbor which makes the number of channels

available for i drop to 0 Finding Critical Neighbor for i

run allocations on {B/bi} (B: set of bids) Ensure truthfulness


VERITAS Truthfulness

• Theorem: VERITAS spectrum auction is truthful, achieves pareto optimal allocations, and runs in polynomial time of O(n3k)

• Proof sketch– Monotone allocationsMonotone allocations: If the bidder wins with bid b,

it also wins with b’ > b when others’ bids are fixed– Critical valueCritical value: Given a bid-set B, a critical value exists

for every allocated bidder– TruthfulnessTruthfulness: If we charge every bidder by its critical

value, no bidder has an incentive to lie


VERITAS Extensions Support various objective functions

VERITAS allocation scheme can sort on broad class of functions of bids

The auctioneer can customize based on its needs

Bidding Formats Range Format: Every bidder i specifies parameter di, and

accepts any number of channels in the range (0, di) Contiguous Format: Bidder requests the channels allocated to

be contiguous


A Closer Look at VERITAS

Revenue curve not monotonically increasing with # of channels auctioned Effect of the pricing scheme Successful auctions require

sufficient level of competition

Enforce competition Choose the proper # of channels

to auction

13

Choosing the number of channels to be auctioned to maximize revenue

Choosing the number of channels to be auctioned to maximize revenue


19



Enabling Trading by Double Auctions

SellersSellers BuyersBuyers

BidsBids

Double Auctions: Sellers and buyers are

bidders Seller’s bid: the minimal price it

requires to sell a channel Buyer’s bid: the maximal price it

is willing to pay for a channel

Auctioneer as the match maker

Select winning buyers and sellers

Winners & Prices


Need Judicious Auction Designs

Bids

SellersSellers BuyersBuyers

Bids

Need to achieve 3 economic properties Budget balance: Payment to

sellers <= Charge to buyers Individual rationality:

Buyer pays less than its bid Seller receives more than its

bid Truthfulness: bid the true

valuation Need to provide efficient

spectrum distribution

$$


Existing Solutions No Longer Apply

Truthfulness

Individual Rationality

Budget Balance

Spectrum Reuse

McAfee’s Double Auction

√ √ √ X

VCG Double Auction √ √ X X

Extension of Single-sided

Truthful Auction

X √ √ √

Our Goal √ √ √ √


Design Guidelines Start from the McAfee design: the most popular truthful

double auction design Achieve all three economic properties without spectrum

reuse

Extend McAfee to assign multiple buyers to each single seller Enable spectrum reuse among buyers

Design the procedure judiciously to maintain the three economic properties


McAfee Double Auctions

Achieve budget balance, truthfulness, individual rationality without spectrum reuse

S1

S2

…

Sk-1

Sk

Sk+1

…

Sm

S1

S2

…

Sk-1

Sk

Sk+1

…

Sm

B1

B2

…

Bk-1

Bk

Bk+1

…

Bn

B1

B2

…

Bk-1

Bk

Bk+1

…

Bn

Sellers’ bidsBuyers’ bids

(k-1) winning buyers, each

paying Bk

≥≥

≥

≤≥

(k-1) winning sellers, each getting paid

Sk

Sacrifice one transaction


Enabling Spectrum Reuse

Map a group of non-conflicting buyers to one seller

Sellers’ bidsBuyers’ bids

S1

S2

…

Sk-1

Sk

Sk+1

…

Sm

S1

S2

…

Sk-1

Sk

Sk+1

…

Sm

B1

B2

…

Bk-1

Bk

Bk+1

…

Bn

B1

B2

…

Bk-1

Bk

Bk+1

…

Bn

Buyer Group G1

Buyer Group G2

Buyer Group G3

≥≥

≥

≤≥


TRUST: Auction Design

Form buyer group

Form buyer group

Bid-independent

Group Formation

Decide the bid of each buyer group;

Apply McAfee

Decide the bid of each buyer group;

Apply McAfee

Buyer group i’s bid = The lowest bid in group i *

#of bidders in group i

Charge individuals in a winning buyer

group

Charge individuals in a winning buyer

group

Uniform pricing within one

winning buyer group

Theorem 1. TRUST is ex-post budget balanced, individual rational, and truthful.


27

Chapter 17 Summary Spectrum is not going to be free (most of it) Economics must be integrated into spectrum

distributions Networking problem: on-demand spectrum allocation Economic problem: truthful (economic-robust) design

Existing solutions fail when enabling spectrum reuse Many ongoing efforts to make this happen in practice


References & Further ReadingsPapers discussed in this chapter: S. Gandhi, C. Buragohain, L. Cao, H. Zheng, and S. Suri, “A general framework for wireless spectrum

auctions,” in Proc. of IEEE DySPAN, 2007. X. Zhou, S. Gandhi, S. Suri, and H. Zheng, “eBay in the sky: Strategy-proof wireless spectrum auctions,”

in Proc. of MobiCom, Sept. 2008. X. Zhou and H. Zheng, “TRUST: A general framework for truthful double spectrum auctions,” in Proc. of

INFOCOM, April 2009.

Further readings: S. Olafsson, B. Glower, and M. Nekovee, “Future management of spectrum,” BT Technology Journal, vol.

25, no. 2, pp. 52–63, 2007. Ofcom, “Spectrum framework review,” June 2004. M. Buddhikot et. al., “Dimsumnet: New directions in wireless networking using coordinated dynamic

spectrum access,” in Proc. of IEEE WoWmoM05, June 2005. T. K. Forde and L. E. Doyle, “A combinatorial clock auction for OFDMA based cognitive wireless

networks,” in Proc. of 3d International Conference on Wireless Pervasive Computing, May 2008. W. Vickery, “Counterspeculation, auctions and competitive sealed tenders,” Journal of Finance, vol. 16,

pp. 8–37, 1961. D. Lehmann, L. O´callaghan, and Y. Shoham, “Truth revelation in approximately efficient combinatorial

auctions,” J. ACM, vol. 49, no. 5, pp. 577–602, 2002. A. Mu’alem and N. Nisan, “Truthful approximation mechanisms for restricted combinatorial auctions:

extended abstract,” in Eighteenth national conference on Artificial intelligence, pp. 379–384, 2002.

28


References & Further Readings R. P. McAfee, “A dominant strategy double auction,” Journal of Economic Theory, vol. 56, pp. 434–450, April 1992. P. Subramanian, H. Gupta, S. R. Das, and M. M. Buddhikot, “Fast spectrum allocation in coordinated dynamic

spectrum access based cellular networks,” in Proc. of IEEE DySPAN, November 2007. Spectrum Bridge Inc., http://www.spectrumbridge.com. P. Subramanian, M. Al-Ayyoub, H. Gupta, S. Das, and M. M. Buddhikot, “Near optimal dynamic spectrum allocation

in cellular networks,” in Proc. Of IEEE DySPAN, 2008. Y. Xing, R. Chandramouli, and C. Cordeiro, “Price dynamics in competitive agile spectrum access markets,” IEEE

Journal on Selected Areas in Communications, vol. 25, no. 3, pp. 613–621, 2007. D. Niyato, E. Hossein, and Z. Han, “Dynamics of multiple-seller and multiple-buyer spectrum trading in cognitive

radio networks: A game theoretic modeling approach,” IEEE Transactions on Mobile Computing, vol. 8, no. 8, pp. 1009–1021, 2009.

V. Rodriguez, K. Mossner, and R. Tafazoli, “Auction-based optimal bidding, pricing and service priorities for multi-rate, multi-class CDMA,” in Proc. Of IEEE PIMRIC, pp. 1850–1854, September 2005.

J. Huang, R. Berry, and M. L. Honig, “Auction-based spectrum sharing,” ACM Mobile Networks and Applications, vol. 11, no. 3, pp. 405–618, 2006.

29

Cognitive Radio Communications and Networks: Principles and Practice By A. M. Wyglinski, M. Nekovee, Y. T. Hou (Elsevier, December 2009) 1 Chapter 17 Auction-based.

Documents

cognitive radio networks

cognitive radio communications

hou elsevier

majority of spectrum

usable spectrum

years of spectrum usage

auctionbased spectrum

portion of spectrum