在本篇論文中，探討在LTE 系統考慮非授權頻譜(Unlicensed band)後，如何解決干擾並提升頻譜使用效率以及頻譜重複使用率 為了要解決叢集小型基地台間的干擾，我提出了基於拍賣理論應用於動態資源分配的想法；有別於傳統的ICIC 演算法雖然簡單易於實做，但頻譜使用並沒有彈性，因此希望利用拍賣競標的概念讓最合適的人去投標取得適合自己使用的資源，讓資源利用最佳化 先以基於第一密封式拍賣方法的拍賣理論提出集中式拍賣演算法，以電信業者為賣方，系統資源(Resource Blocks)為拍賣商品，小型基地台(Smallcells)為投標者，以編號最小的基地台當任拍賣官，來蒐集其他基地台投標資訊，並分配完資源後通知各個基地台所分配到的RB；出價原則為讓可得到更高throughput 的基地台有較大的優勢去贏得競價，進而提升整體系統效能；然而，在集中式拍賣演算法中出現公平性以及決定權於單一基地台的問題，我們提出改良後的分散式拍賣演算法，並使用英式拍賣方法為基礎做拍賣理論的動態資源分配，同時取消掉拍賣仲裁者角色，以來回出價的概念進行競拍，並在投標者的出價原則中加入了負載持有平衡比例的想法，改善集中式拍賣演算法中的投標者滿意度問題;In this research with the long term evolution (LTE) in the unlicensed band(LTE-U), we discuss how to eliminate the interference for the cluster of small cells, increase the throughput, and improve the frequency reuse factor is a issue in recently years.This thesis will present the dynamic resource allocation based on the auction theory for eliminating the interference for a cluster of small cells. This solution is different from the ICIC, which is easy to be implemented but its resource allocation method is not flexible. The proposed method of dynamic resource allocation based on the auction theory lets the base station get the appropriate resource and makes the use of resource more efficiently. In the first, we propose an algorithm based on sealed-bid first price auction – central auction algorithm. In this auction algorithm, the operators are the sellers and the smallcells are the buyers and small cells bid for the auctions (system resource blocks) based on their channel quality. The smallest number of small cell is the auctioneer and it will collect the bidding of other smalcells and inform them that which resource blocks they can use. The bidding principle is to let the small cells which can modulate better throughput get the resource blocks more easily. That’s why system throughput in auction algorithm gets higher performance. However, the central auction theory has the fairness problem that some small cells may not grasp any resource during the auction. This problem cause the reason that we present an improved auction theory which is based on the English auction - distributed auction theory. In this algorithm, we remove the auctioneer and use the bidding round way to auction. For solving the fairness problem, this distributed auction algorithm includes the concept of holding fairness to improve it.
