A prior-free revenue maximizing auction for secondary spectrum access
A prior-free revenue maximizing auction for secondary spectrum access
复制标题
DOI：
10.1109/infcom.2011.5935311
复制
发表时间：
2011-04
期刊：
影响因子：
--
通讯作者：
A. Gopinathan;Zongpeng Li
中科院分区：
文献类型：
--
作者：
A. Gopinathan;Zongpeng Li
动态频谱分配已被证明有希望缓解频谱稀缺问题。在这种模式中，主要用户短期租赁未充分利用的频谱块给次要用户。主要用户可能需要财务动机来共享频谱，因为他们承担获得频谱许可证的成本。拍卖是一种自然的创收机制。最近的频谱拍卖设计强烈假设主用户知道用户估值的概率分布。我们研究收入最大化的频谱拍卖在更现实的事先免费设置，当用户估值的信息是不可用的。构建了一个两阶段拍卖框架。在第一阶段，我们设计了一个策略验证机制，计算一个子集的用户与干扰的频谱分配，使潜在的收入在第二阶段最大化。量身定制的支付方案确保了这一阶段的真实投标。第二阶段，我们设计了一个随机竞争拍卖，并通过投标独立性证明了它的防策略性。采用概率技术，我们证明，我们的拍卖产生的收入，至少是1比3的最佳收入，提高了最好的已知比例1比4证明类似的设置。
Dynamic spectrum allocation has proven promising for mitigating the spectrum scarcity problem. In this model, primary users lease chunks of under-utilized spectrum to secondary users, on a short-term basis. Primary users may need financial motivations to share spectrum, since they assume costs in obtaining spectrum licenses. Auctions are a natural revenue generating mechanism to apply. Recent design on spectrum auctions make the strong assumption that the primary user knows the probability distribution of user valuations. We study revenue-maximizing spectrum auctions in the more realistic prior-free setting, when information on user valuations is unavailable. A two-phase auction framework is constructed. In phase one, we design a strategyproof mechanism that computes a subset of users with an interference-free spectrum allocation, such that the potential revenue in the second phase is maximized. A tailored payment scheme ensures truthful bidding at this stage. The selected users then participate in phase two, where we design a randomized competitive auction and prove its strategyproofness through the argument of bid independence. Employing probabilistic techniques, we prove that our auction generates a revenue that is at least 1 over 3 of the optimal revenue, improving the best known ratio of 1 over 4 proven for similar settings.