Dynamic Spectrum Auction Based on Coexistent Matrix

Dynamic spectrum auction is an effective way to stimulate primary users to lease their idle spectrum and meanwhile solve the competitions among secondary users through bidding. Due to the spatial reusability of spectrum, multiple users which are separate enough can have access to the same spectrum simultaneously without interfering each other. In this paper, we propose a spectrum auction framework based on physical interference model. Instead of conflict graph we propose coexistent matrix which can characterize the cumulative interference effect to achieve a reliable allocation. A third party interference management institute is employed to compute the coexistent matrix in our protocol. Based on coexistent matrix, we study the truthful rules for our spectrum auction. We design our spectrum auction protocol to make bidders hard to form a collusive group. For the third party institute, an algorithm which can generate the coexistent matrix with polynomial time complexity is also presented. Numerical experiments are employed to evaluate the performance of our spectrum auction.