行動流量近年來增長得非常快速，尤其是針對熱門影片的需求。大型基地台與小型基地台結合的異構網路也被用來因應行動網路流量的增長。由於回程網路的流量限制，為了避免流量過多，影片暫存機制將能降低網路流量。以學習為基礎的影片暫存技術是以使用者的喜好不會在一朝一夕所改變當作立足點，本論文的目標是提出一個計畫能在使用最少回程網路流量的情況下，服務熱門影片給用戶。
在學習的階段，有相同喜好的用戶會被聚在同一群並分配給小型基地台，每群用戶的特性會被整合成其對應到的小型基地台的特性，接著會藉由盧比變化碼將影片編碼的方式，讓多個小型基地台有合作的關係，此舉能共享儲存裝置的容量。最後，在服務階段，對於新加入的用戶，會依照他們要求的影片，和小型基地台的特性相比，決定要把新用戶分到哪一個小型基地台。另外，將用戶分群也可以用在另外一個應用上:當多個距離不遠的用戶同時想看某一段影片時，藉由共享下載量的方式，也可以減少回程網路的流量。

Mobile traffic has grown fast in recent years, particularly for delivering popular video clips. Based on a combination of Macro Cells and various Small Cell (SC) technologies, Hyper-dense Heterogeneous Networks (HetNets) is gaining increasing attention due to huge demand and popularity of mobile video traffic. To avoid bottleneck in the limited capacity of backhaul link to the core network, caching in the network edge in such a way that the buffered video can be delivered with less network latency and traffic load is very promising. A learning-based caching plan, on the basis of users’ preferences may not subject to change rapidly, is proposed in the thesis. Our goal is to build a distributed caching plan for serving popular video clips over HetNets with least possible backhaul traffic.
In the learning phase of the proposed learning-based method, cluster and assign similar users to SCs using spectral clustering first. Second aggregate the users’ requests to be the request profile of the corresponding SCs. Third, share the caching space among cooperated SCs with the help of distributed LT codes. During the serving phase, new coming users will be assigned to appropriate SCs based on similarity between users and SCs. However, the assignment of old users remains the same on the assumption that users’ preferences do not subject to change rapidly. Moreover, the clustering information can be applied to another application that several users can share downloading cooperatively if a group of smartphone users request watching the same video clips almost the same time, subsequent decreases the load of backhaul bandwidth.

[1]. Forecast, Cisco VNI, “Cisco Visual Networking Index: Global Mobile Data Traffic Forecast Update, 2015–2020 White Paper,” Cisco Public Information, February 2016.
[2]. Insoo Hwang, Bongyong Song, and Samir S. Soliman, “A Holistic View on Hyper-Dense Heterogeneous and Small Cell Networks,” IEEE Communications Magazine, June 2013, pp. 20-27.
[3]. M Luby, “LT Codes,” in Proceedings of the 43rd Annual IEEE Symposium on Foundations of Computer Science, November 2002, pp. 271–280.
[4]. Maatouk, Ghid, and Amin Shokrollahi, “Analysis of the Second Moment of the LT Decoder,” IEEE Transactions on Information Theory, May 2012, pp. 2558-2569.
[5]. J.W. Byers, M. Luby, M. Mitzenmacher, and A. Rege, “A Digital Fountain Approach to Reliable Distribution of Bulk Data,” Proceedings of ACM SIGCOMM ‘98, Vancouver, September 1998, pp. 56-67.
[6]. Keller, Lorenzo, et al, “MicroCast: Cooperative Video Streaming on Smartphones,” Proceedings of the 10th international conference on Mobile systems, applications, and services. ACM, June 2012, pp. 57-70.
[7]. Golrezaei, N., Shanmugam, K., Dimakis, A. G., Molisch, A. F., & Caire, G, “Femtocaching: Wireless Video Content Delivery through Distributed Caching Helpers,” in Proceedings IEEE INFOCOM, March 2012, pp. 1107-1115.
[8]. Shanmugam, K., Golrezaei, N., Dimakis, A. G., Molisch, A. F., & Caire, G. “Femtocaching: Wireless Content Delivery through Distributed Caching Helpers,” IEEE Transactions on Information Theory, December 2013, pp. 8402-8413.
[9]. M. S. ElBamby, M. Bennis, W. Saad, and M. Latva-aho, “Content-aware user clustering and caching in wireless small cell networks,” in IEEE Intl. Symp. on Wireless Communications Systems (ISWCS), Barcelona, Spain, August 2014, pp. 945–949.
[10]. WANG, Zheshen, et al, “Youtubecat: Learning to categorize wild web videos.” In Computer Vision and Pattern Recognition (CVPR), 2010 IEEE Conference on. IEEE, 2010. pp. 879-886.
[11]. U. Gargi, W. Lu, V. Mirrokni, and S. Yoon, “Large-scale community detection on YouTube for topic discovery and exploration,” In Proceedings of Fifth International AAAI Conference on Weblogs and Social Media, July 2011.
[12]. Yang, Jaewon, and Jure Leskovec, ”Overlapping community detection at scale: a nonnegative matrix factorization approach,” Proceedings of the sixth ACM international conference on Web search and data mining, February 2013, pp. 587-596.
[13]. Ng, Andrew Y., Michael I. Jordan, and Yair Weiss, “On spectral clustering: Analysis and an algorithm,” Advances in neural information processing systems, 2002, pp. 849-856.
[14]. Newman, Mark EJ. “Modularity and community structure in networks,” Proceedings of the national academy of sciences, 2006, pp. 8577-8582.
[15]. P. Gill, M. Arlitt, Z. Li, and A. Mahanti, “Youtube Traffic Characterization: A View From the Edge,” In Proceedings of the ACM SIGCOMM conference on Internet Measurement (IMC), 2007, pp.15-28.
[16]. M. Zink, K. Suh, Y. Gu, and J. Kurose, “Characteristics of youtube network traffic at a campus network-measurements, models, and implications,” Computer Networks, 2009, pp.501 − 514.