摘要
由於大多數的傳送端裝置都是以電池提供能量，因此提升能量使用效率可以 同時提升成本效益與延長電池使用壽命。因為通道會隨著時間以及使用者的位置 不同而變化，傳送端可以利用這些變化所造成的分集(diversity)，視不同時間 之通道增益適當地調整其傳輸能量，進而大大的提高其能量的使用效率。 在本論文中我們將會探討如何在時域上有效地分配傳輸能量以達到總能量 消耗最小化。在已知當前通道增益與未來通道的統計特性，但不知未來通道增益 資訊的情況之下，我們將設計一排程器既能節省能源又能滿足其服務質量 (Quality of Service)，即資料傳輸流量及限時條件。為了探討通道衰減、限時 條件與當前通道增益之間的相互影響，我們在此只討論一基本的排程問題，即我 們只討論抵達時間間格固定的單一封包，如:VoIP 網路電話，影音串流都是這類 型的限時通訊方式，另外，我們沒有考慮中斷機率(Outage probability)的發生。 利用反住水(Inverse Water-Filling)理論，我們可以推導出最佳非因果 (non-causal)的排程器，而在只有兩個時段(time slot)的情況下，也可推導出 最佳的因果排程器。此外，我們利用動態規劃法(dynamic programming)也可推 導出最佳分配法，但在總時段數大於二時沒有封閉式解法(Closed-form)，即無 法用基本函數表示。於是我們提出了兩個次最佳排程方案，一個利用了中央極限 定理(Central Limit Theorem)，另一個使用了反住水理論。兩個方法皆是由通 道意識與延遲意識所組成的線性組合，且模擬結果顯示當傳送量大時，兩的方法 皆接近於最佳排程結果。 此外，我們推廣到多使用者的例子。在第一階段我們只有使用者們的通道特 性。在限時條件下，使用者們的分配順序並不重要，因此在此階段只需決定該分 配多少時段給每個使用者。之後我們將可把問題視為多個獨立的單一使用者、單 載波問題。在多個使用者情況下，排程問題可以分成資源分配、通道指定，即傳 送方必須利用順序統計法(order statistics)決定該通道應當指派給哪位使用 者，並決定該傳送多少能量以達成總能量消耗最小化的目的。
並列摘要
For many wireless transmitters, since most devices are battery powered, increased energy e±ciency in data transmission provides signi¯cant bene¯ts. Higher energy e±- ciency may result in prolonging the lifetime of the battery. We seek to ¯nd an energy-saving scheduler that sends a packet of R bits within a hard delay deadline K over fading channels. The scheduling policies needs to determine the number of bits transmitted in the current time slot with only the knowledge of current channel state information and the channel statistics of the future channel while satisfying the quality of service QOS constraints as the deadline expired in order to minimize the total energy consump- tion. In this thesis, we will focus on the interaction between fading, hard deadlines, and causal channel information by studying transmission of only a single packet, and thus do not consider random arrivals since there are applications with deterministic packet arrivals, i.e., VoIP or video streaming where packets arrive regularly and each must be received within a short delay window. Although it is more reasonable to consider ran- dom arrivals and outage probability that allows few packets missing, to emphasize the relationship between fading, hard deadlines, and causal channel information, we only consider a fundamental scheduling problem that one packet and no outage is allowed. An optimal non-causal scheduling policy is derived by inverse water-¯lling (IWF) method and an optimal causal scheduling policy is also derived for total time slots K = 2. We also develop a dynamic programming formulation that leads to an optimal transmission schedule, however, it is hard to express as a closed form when K > 2. Thus, we propose two suboptimal scheduler which give simple structure for general problems, and one utilizes central limit theorem (CLT) for approximation while the other is inspired by the IWF method. The policies are composed of a linear combination of channel-awareness term and delay-awareness term. The numerical results show that the proposed policies are nearly optimal when R is large. In addition, we extend our work to multiple user case. At the ¯rst phase, we only have the channel pdfs of users. Since delay constraint speci¯es only that the rate is achieved in K blocks, the order of which the users are scheduled within the K blocks is not important in this phase, so the scheduling boils down to sorting out the number of blocks being allocated to each user. After deciding the number of blocks allocated to each user, the problem can be treated as independent single-user single-carrier problems with competitions. With multiple users, the scheduling problem is composed of distributing resources, channel assignment that the transmitter requires to decide which user occupies the channel at any given time slot by order statistic method and bit allocation that how many bits should be allocated in order to minimize the total energy consumption of all users.
參考文獻
延伸閱讀
- 陳立騫（2002）。有時間時窗限制與加權提早/延遲成本之單機排程研究〔碩士論文，中原大學〕。華藝線上圖書館。https://doi.org/10.6840/cycu200200423
- 徐浩均（2004）。雙階段流程型工廠在有限暫存區限制下之排程問題〔碩士論文，中原大學〕。華藝線上圖書館。https://doi.org/10.6840/cycu200400463
- Chang, P. C., & Chung, Y. K. (2004). 加權完工時間最小化之動態排程問題. 工業工程學刊, 21(6), 567-575. https://doi.org/10.29977/JCIIE.200411.0006
- Yu, H. F. (2013). Single-Channel Data Broadcasting under Small Waiting Latency. Journal of Applied Mathematics, 2013(), 299-306-534. https://doi.org/10.1155/2013/629350
- 鍾國言（2017）。Dishing up Scheduling Problem with Capacity Constraint〔碩士論文，朝陽科技大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0078-2712201714432321