Multi-seller combinatorial spectrum auction with reserve prices

Abstract

Efficient spectrum allocation is becoming more and more important in wireless networking. Auction is believed to be an effective way to address the problem of spectrum shortage, by dynamically redistributing spare channels among service providers. Combinatorial auction gives buyers the freedom to place bids on combinations of channels rather than individual channels. For example, continuous channels are easier to operate on and more valuable to buyers. Unlike conventional commodities, spectrum features spatial reusability, which depends on various transmission ranges of heterogeneous channels, making combinatorial spectrum auction more challenging. Existing works on combinatorial spectrum auction only consider a single seller who tries to minimize her cost. However, in multi-seller markets, each seller has a reserve price, below which the seller is reluctant to sell her channel. In this paper, we propose a combinatorial auction mechanism for multiple sellers with specified reserve prices. We design an efficient greedy algorithm to determine auction winners by the average virtual bids, which is decided by buyers' real bids and the reserve prices of channels. Simulation results show that our proposed auction mechanism can achieve higher social welfare than existing auction mechanisms without reserve prices.