在本文中，我們考慮了無線感知網路系統模型組成為一個頻譜擁有者 (PO)、多個主要使用者 (PU) 和多個次要使用者 (SU)。我們設計了一個拍賣模式的頻段分配，其中 SU 作為競標者來競標頻段，而 PO 作為拍賣者拍賣空白頻段賺取利潤。
現今討論頻譜分論文都假設所有頻段皆為相同，但是本篇論文考慮更實際情況的異質性頻段。此外，我們允許每個 SU 可以公布他的喜好類型向量和決定不同的需求數量。本篇論文中提出了線性規劃關係和全付機制來進行有效率的頻譜分配，其中考量了經濟屬性的限制和預算限制條件，在文中有正式的說明與分析。最後，數值模擬分析的結果在 SU 的效能函數、PO 的收益函數和社會福利函數都有較優的結果，無論是跟同質性頻段比較或者是與其他參考文獻所提出的方法比較。 Cognitive radio network is a novel commutation paradigm in recent years which can significantly improve the spectrum utilization. In this paper, we consider a cognitive radio network model consisting of a primary owner (PO), multiple primary users (PU) and multiple secondary users (SU). We design an auction-based spectrum allocation mechanism where the SUs bid to purchase the spectrum bands from the PO who acts as the auctioneer. Existing auction-based spectrum allocation mechanisms mostly assume the spectrum bands are homogenous. However, we consider the heterogeneous spectrum bands. Moreover, we allow the SUs can announce their type and determine their different demands. We propose the Linear Programming Relaxation and All-Pay Mechanism to achieve the high efficiency of the spectrum allocation which involves the economic properties and budget constraints . Finally, the numerical results show performance improvements in terms of SUs' utilities, PO's revenue and social welfare.
