Truthful Auction Mechanisms For Spectrum Allocation In Cog-Nitive Radio Networks

Radio spectrum is a critical and scare resource for wireless communications. In recent years, with the explosion of novel wireless devices and applications, the in-creasing growth of spectrum requirement has outpaced available spectrum resources. Cognitive Radio Networks (CRNs) has emerged as a promising solution to address the above dilemma, where licensed users (primary users) can open up their idle spec-trum to unlicensed users (secondary users). In this situation, a critical issue is how to efficiently address spectrum sharing among licensed users and unlicensed users.Auctions are the well recognized and de facto approaches for redistributing spec-trum in CRNs that can achieve both fairness and allocation efficiency. In this way, licensed users can gain financial benefits by leasing their idle spectrum to unlicensed users to access, i.e., a win-win situation. The key challenge of designing efficient spec-trum auction is to provide truthfulness while exploiting spectrum reusability. Truth-fulness is a critical property required to stimulates bidders to reveal their valuations of spectrum truthfully. Spectrum reusability indicates that a channel can be allocated to multiple well-separated users, and thus can highly improve spectrum utilization. The spectrum reusability always destroys the truthfulness of traditional auctions, and also makes it hard to design efficient spectrum allocation algorithm.Although a number of spectrum auction designs have been proposed in recent years, there still exists a few issues need to be solved. First, most of works are only for the two-step allocation scenario where bidders can only compete for the secondary access right of channels, while users may compete for the access rights according to their QoS demands in single-step allocation. However, existing auction designs for single-step allocation do not exploit spectrum reusability. Second, current works only focus on truthfulness, and ignored the potential threats where bidders may use multiple false names to bid. Finally, current works provide strong truthfulness while ignored the auction efficiency. To address the above issues, we propose efficient spectrum auction mechanisms correspondingly. In details, the main contributions of this dissertation can be summarized as follows.(1) We investigate the single-step allocation problem and propose a QoS-aware truthful auction framework TRUMP for large-scale networks. In the single-step al-location scenario, the request on spectrum access rights of a user relies on its QoS demands. Existing auction designs for single-step allocation do not exploit spectrum reusability. To solve this problem, we propose TRUMP, a truthful spectrum auction considering QoS demands and spatial reuse. TRUMP is the first mechanism which exploits the spectrum reusability for single-step allocation. The theoretical analysis proves that TRUMP achieves truthfulness and with polynomial-time complexity and extensive simulation results show that TRUMP outperforms previous works in terms of spectrum utilization and social welfare.(2) We investigate the impact of false-name bids in dynamic spectrum auctions and propose a false-name-proof auction framework for large-scale networks. A us-er equipped with cognitive radio can easily generate multiple "names". In this way, auction manipulation by the bidders can go beyond the cheating with false valuations and bidders can submit false-name-bids. We first show that false-name bid cheating is easy to form in spectrum auction and can drastically reduce the revenue of the auction-eer. To solve this issue, we propose ALETHEIA, a novel flexible, false-name-proof auction framework for large-scale networks. ALETHEIA not only guarantees truthful-ness but also resists false-name bids. It provides the bidders the flexibility of diverse demand formats and incurs low computational overhead. Simulation results show that ALETHEIA achieves both high spectrum redistribution efficiency and auction efficien-cy.(3) We investigate the secondary spectrum auction designs and propose approxi-mately truthful spectrum auction mechanisms for large-scale networks. When consid-ering spectrum reusability, truthful auction design typically involves solving NP-hard problems. Previous work always guarantee the truthfulness while ignoring the auction efficiency. We solve the problem in a novel perspective by relaxing the constraint of ensuring strong truthfulness, to achieve a trade-off between allocation efficiency and truthfulness. Specifically, we first propose a computationally efficient mechanism that achieves truthful in expectation, which guarantees any bidder cannot gain positive prof-it in expectation from untruthful bid. Then we propose a hard-to-manipulate auction which makes it hard to manipulate the auction for any profit gains. Simulation results show that our mechanisms can achieve significant improvement over the state-of-the-art mechanisms.