Short-interval Auction Design for Dynamic Spectrum Access in a Cell of Single Technology
Date Issued
2009
Date
2009
Author(s)
Liu, Hao-Huai
Abstract
With the technological advances in mobile communications and spectrum management and the divergent service demands of end-users, Dynamic Spectrum Access (DSA) has been recently proposed to increase bandwidth usage efficiency and provide innovative or specialized services such as low-cost telecommunication services with low QoS guarantee through spectrum sharing and/or trading. In the proposed DSA paradigm, we focus on the cooperative sharing of access rights where a secondary user (ex: end-user of a Mobile Virtual Network Operator, MVNO) has access to a licensed band with the permission from its primary right holder (ex: a Mobile Network Operator, MNO). Such cooperative sharing may facilitate many innovative business models or trading mechanism designs and new applications such as business-specific or real-time sports telecommunication services provided by the MVNO. In this thesis, we investigate the design of a short-interval market because the under-utilized spectrum bands often appear in short duration ranging from tens of seconds to tens of minutes according to many empirical reports. Because auction is widely viewed as an efficient way of allocating scarce resources [GBC07]and has been widely adopted in spectrum market, an auction market is designed, which consists of three parties: MNO, Spectrum Broker, and MVNO. In this design of short-interval auction market, the End-users access the MNO’s base station (BS) by using the spectrum that MVNO leases from MNO through the brokerage of a SB. The market design objective is to design a platform which provides MNO and MVNOs with sufficient incentives to trade and thus increase the spectrum utilization and stimulate the growth of innovative services. We study a 1-MNO-N-MVNO market. In view that there are multiple units of homogeneous objects to trade and little time for trading, we refer to the auction proposed by Engelbrecht and Kahn in 1998: a pay-as-bid, single-round sealed-bid auction, bidder submits Ji bids if he has Ji demands, and the bid selection policy is allocating K objects to the most highest K bids. This auction is not suitable for DSA market because in DSA market the value of channels partly depend on the amount of channels allocated. Instead, a demand responsive bidding form where MVNO has to represent the bid price of each number of channels is used. The advantage of it is there will not be asymmetry of the channel resource value and the payment. Furthermore, the reserve price and bid increment are also set to enhance auction efficiency and secure MNO’s profit. We then design the trading protocol and message passing for the market. We first mathematically formulate Spectrum broker’s bid selection problem as the SB maximizes the sales revenue by allocating channels under the entrance, bid and total channel availability constraints. The bid decision problem is formulated as a problem where each MVNO view other MVNOs’ bidding behaviors as a single opponent’s and also aggregate their bids to decide the bid offer which maximizes the overall expected profit under the constraints of entrance, bid and profit. The challenge of solving the auction problem lies in the complicated computation of winning probabilities of different bid offers. e then explore the properties of MVNO’s bidding strategy and the auction outcome by considering a simple example of N=2. There are a business MVNO with a 3-channel demand and a discount MVNO with a 4-channel demand, and the MNO offers 5 channels for the two to bid. A heuristic MVNO behavior assumption of “Concentrated bidding strategy” is assumed, where considering the SB’s total sales maximization strategy, an MVNO would bid high on getting the demanded number of channels and bid minimum on number of channels lower than demanded to increase the probability of winning the demanded. The analyses of this example are as follows:. BDPi model gives MVNO a higher expected profit than that of the heuristic strategy.. Bidding strategy is sensitive to the penalty of unsatisfied guaranteed QoS.MVNO with high penalty rate will bid the highest on the demanded to acquire all demanded channels while MVNO with low penalty rate will bid the lowest on channels lower than the demanded to obtain the highest profit.. Both MNO and MVNOs gain profits.In this example, demand is higher than supply and the proposed auction design sells out all channels. Also, MNO’s profit is 16-fold the reserve price, and the two MVNOs could earn profit via auction.ey words - DSA, secondary market, cooperative sharing, MNO, MVNO, spectrum broker, value chain, business model, market structure, auction, short-interval, market mechanism
Subjects
DSA
secondary market
cooperative sharing
MNO
MVNO, spectrum broker, value chain, business model
market structure
auction
short-interval
market mechanism
Type
thesis
File(s)![Thumbnail Image]()
Loading...
Name
ntu-98-R96546024-1.pdf
Size
23.53 KB
Format
Adobe PDF
Checksum
(MD5):d41cbea8c6ee840789a1ad73c55421b8