Licensed shared access for 5G: Which auction mechanism to choose?

Abstract

Licensed Shared Access (LSA) is a complementary solution allowing Mobile Network Operators (MNOs) to use another incumbent's frequency spectrum after obtaining a proper license from the regulator.Using auctions to allocate those LSA-type licenses is a natural approach toward an efficient use of spectrum, by controlling the incentives for MNOs to declare their true valuation for the spectrum and allocating it to those who value it the most. A specificity of LSA licenses lies in the interactions among buyers, due to possibly overlapping coverage areas, this allows for allocating the same spectrum to several MNOs.In this paper, we review the existing mechanisms taking into account such radio interference constraints, propose new ones, and compare their performance. We show how to increase the revenue, while maintaining truthful-telling, of all-or-nothing auction mechanisms by introducing a reserve price per bidder. We also investigate extensions of those mechanisms, namely when the management of interference among base stations is more subtle than partitioning base stations into groups of non-interfering base stations. For each mechanism, we show how to optimize a trade-off between expected fairness, expected revenue and expected efficiency by carefully working with groups and reserve prices. Simulations suggests that the extension of those mechanisms may lead to increase an indicator combining allocation fairness, social welfare and seller's revenue by more than 20%.

Multi-Block Ascending Auctions for Effective 5G Licensed Shared Access

“…In the second version, we remove those representatives so that the auction will be between the auctioneer and bidders. This is an extension of our work in [8].…”

“…The difference between those two quantities is then (C VCG h,n − j∈g h \i b j,n ). We now prove that the proposed mechanism with allocation given in (8) and payment given in ( 9) is truthful. Let us first introduce the two following lemmas.…”

“…Authors of LSAA claim that LSAA is truthful. However, it is not the case as we have shown in [6], and in [27] we have proposed two variants that are truthful: TLSAA and TLSAA2. Both mechanisms compute the group-bid of each group in the same way as LSAA, however TLSAA is more efficient than LSAA, with a lower revenue.…”

“…LSAA [7] is the first auction mechanism which was proposed as a candidate for the LSA context, it performs well in terms of social welfare assuming truthtelling by bidders, but sincere bidding is not an optimal strategy for bidders. In [8], we modified the payment rule of LSAA to make it truthful. For those works there are two limitations:…”

“…Proposition 1. For the proposed mechanism with allocation rule in (8) and payment rule in (9), truthful bidding, is a dominant strategy i.e., for each bidder i, proposing (b i,1 , .., b i,K ) = (v i,n , .., v i,K ) maximizes its utility.…”