【摘要】：隨著蜂窩數(shù)據(jù)業(yè)務(wù)和流量迅速增長,蜂窩網(wǎng)絡(luò)面臨的數(shù)據(jù)流量壓力越來越大。受網(wǎng)絡(luò)資源的限制,傳統(tǒng)的增加蜂窩系統(tǒng)容量的方法如縮小小區(qū)范圍、增加可用頻譜等,越發(fā)難以為繼。如何對(duì)蜂窩網(wǎng)絡(luò)系統(tǒng)進(jìn)行流量卸載成為學(xué)界和業(yè)界關(guān)注的重點(diǎn)。D2D通信技術(shù)支持兩個(gè)終端用戶直接進(jìn)行通信,相比于傳統(tǒng)的蜂窩通信,D2D通信可以減輕蜂窩基站的負(fù)擔(dān)。此外,邊緣緩存是實(shí)現(xiàn)流量本地化、進(jìn)行流量卸載的有效方法。將D2D通信與緩存相結(jié)合,使得蜂窩網(wǎng)絡(luò)中緩存在終端的數(shù)據(jù)不僅可以滿足終端自身需求,也可以通過D2D通信與鄰近終端共享,這樣將大大減輕網(wǎng)絡(luò)流量負(fù)載。因此研究蜂窩系統(tǒng)中基于D2D通信的終端緩存策略以最大化系統(tǒng)流量卸載性能具有重要的理論意義和實(shí)用價(jià)值。首先,本論文對(duì)蜂窩系統(tǒng)中基于D2D通信的終端被動(dòng)緩存的流量卸載性能進(jìn)行了分析。當(dāng)終端緩存有文件時(shí),終端的文件請(qǐng)求可通過自身緩存或者基于終端間D2D通信的文件共享滿足。本論文對(duì)兩種不同的終端移動(dòng)場(chǎng)景下終端緩存的流量卸載性能進(jìn)行了研究。在靜態(tài)網(wǎng)絡(luò)中,終端間的拓?fù)潢P(guān)系可視作是穩(wěn)定的,因此可以對(duì)各D2D鏈路的通信進(jìn)行調(diào)度,但D2D鏈路的傳輸需求在不同時(shí)刻是變化的,本論文基于最大加權(quán)獨(dú)立集模型提出了一個(gè)啟發(fā)式算法來獲得系統(tǒng)流量卸載比率。在高移動(dòng)性網(wǎng)絡(luò)中,終端間不存在穩(wěn)定的拓?fù)潢P(guān)系,本論文將移動(dòng)終端位置建模為泊松點(diǎn)過程,并通過隨機(jī)幾何分析得到系統(tǒng)流量卸載比率。接下來,本論文研究了蜂窩系統(tǒng)中基于D2D通信的終端主動(dòng)緩存策略。終端可調(diào)整自身緩存內(nèi)容以最大化系統(tǒng)流量卸載量�？紤]到移動(dòng)終端有不同的移動(dòng)性和緩存空間、需要緩存的文件有不同的大小和流行程度這些因素,問題建模為一個(gè)組合優(yōu)化問題并被證明是NP難題。由于終端和文件的數(shù)目較大,通過指數(shù)級(jí)復(fù)雜度獲得最優(yōu)解并不實(shí)際。本論文提出了一個(gè)低復(fù)雜度分布式終端間緩存調(diào)整算法來實(shí)現(xiàn)移動(dòng)終端相遇時(shí)的緩存內(nèi)容調(diào)整,并理論上推導(dǎo)了該算法能達(dá)到的性能界�；诮K端間緩存調(diào)整算法,本論文提出了需要基站提供輔助信息的基站輔助流量卸載算法,為了減少基站輔助帶來的控制信息負(fù)荷,進(jìn)一步提出了由終端進(jìn)行參數(shù)估計(jì)的完全分布式流量卸載算法。最后,本論文對(duì)基于D2D協(xié)作傳輸?shù)膫鬏敳呗院徒K端緩存策略進(jìn)行了聯(lián)合優(yōu)化設(shè)計(jì)。多個(gè)終端可以以D2D協(xié)作傳輸?shù)姆绞较蛲唤K端傳輸同一文件以提高通信質(zhì)量,對(duì)抗衰落,結(jié)合終端緩存策略進(jìn)一步系統(tǒng)提升流量卸載性能。本論文將協(xié)作傳輸策略和終端緩存策略的聯(lián)合優(yōu)化問題分解為求解傳輸策略的子問題和求解緩存策略的主問題。本論文將傳輸策略子問題轉(zhuǎn)化為有限步MDP問題,給出了該MDP問題貝爾曼最優(yōu)性方程,通過求解.貝爾曼方程提出了系統(tǒng)傳輸策略進(jìn)行鏈路選擇;得到系統(tǒng)傳輸策略后,以梯度投影法為工具,對(duì)緩存策略主問題進(jìn)行求解,得到了系統(tǒng)緩存策略。
[Abstract]:With the rapid growth of cellular data services and traffic, the pressure of data traffic on cellular networks is increasing. Due to the limitation of network resources, the traditional methods to increase the capacity of cellular systems, such as narrowing the cell size and increasing the available spectrum, are becoming more and more difficult to sustain. In addition, edge caching is an effective way to localize traffic and unload traffic. By combining D2D communication with caching, the data cached at the end of the cellular network is cached. It can not only meet the needs of the terminal itself, but also share with the adjacent terminals through D2D communication, which will greatly reduce the network traffic load. Therefore, it is of great theoretical significance and practical value to study the terminal buffer strategy based on D2D communication in cellular system to maximize the system traffic unloading performance. The unloading performance of terminal passive cache based on D2D communication is analyzed. When the terminal caches a file, the file request of the terminal can be satisfied by its own cache or file sharing based on D2D communication between terminals. In the state network, the topological relationship between terminals can be regarded as stable, so the communication of each D2D link can be scheduled. But the transmission demand of the D2D link varies at different times. Based on the maximum weighted independent set model, a heuristic algorithm is proposed to obtain the system traffic unload ratio. There is no stable topological relationship between terminals. In this paper, the location of mobile terminals is modeled as a Poisson point process, and the unloading rate of system traffic is obtained by stochastic geometric analysis. Next, the terminal active caching strategy based on D2D communication in cellular systems is studied. Considering that the mobile terminal has different mobility and cache space, the size and popularity of the files that need to be cached are different, the problem is modeled as a combinatorial optimization problem and proved to be a NP-hard problem. A low-complexity distributed inter-terminal cache adjustment algorithm is proposed to adjust the cache content when mobile terminals encounter each other, and the performance bound of the algorithm is deduced theoretically. A fully distributed traffic uninstallation algorithm based on terminal parameter estimation is proposed. Finally, a joint optimization design of transmission strategy and terminal caching strategy based on D2D cooperative transmission is carried out. Multiple terminals can transfer the same file to the same terminal by D2D cooperative transmission. This paper decomposes the joint optimization problem of cooperative transmission strategy and terminal cache strategy into the sub-problem of solving transmission strategy and the main problem of solving cache strategy. The Bellman optimality equation of the MDP problem is given, and the system transmission strategy is proposed for link selection by solving the Bellman equation.
【學(xué)位授予單位】：浙江大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TN929.5
，

本文編號(hào)：2216600