Online Content Popularity Prediction and Learning in Wireless Edge Caching

被引:63
作者
Garg, Navneet [1 ]
Sellathurai, Mathini [1 ]
Bhatia, Vimal [2 ]
Bharath, B. N. [3 ]
Ratnarajah, Tharmalingam [4 ]
机构
[1] Heriot Watt Univ, Sch Engn & Phys Sci, Inst Sensors Signals & Syst, Edinburgh EH14 4AS, Midlothian, Scotland
[2] Indian Inst Technol Indore, Elect Engn Dept, Indore 453552, India
[3] Indian Inst Technol Dharwad, Elect Engn Dept, Dharwad 580011, Karnataka, India
[4] Univ Edinburgh, Inst Digital Commun, Edinburgh EH8 9YL, Midlothian, Scotland
基金
英国工程与自然科学研究理事会;
关键词
Predictive models; Base stations; Computational modeling; Analytical models; Wireless communication; Computational complexity; Minimization; Linear prediction; caching; Poisson point process (PPP); online learning; CONTENT PLACEMENT; NETWORKS; DELIVERY;
D O I
10.1109/TCOMM.2019.2956041
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
Caching popular contents in advance is an important technique to achieve low latency and reduce the backhaul costs in future wireless communications. Considering a network with base stations distributed as a Poisson point process, optimal content placement caching probabilities are obtained to maximize the average success probability (ASP) for a known content popularity (CP) profile, which in practice is time-varying and unknown in advance. In this paper, we first propose two online prediction (OP) methods for forecasting CP viz., popularity prediction model (PPM) and Grassmannian prediction model (GPM), where the unconstrained coefficients for linear prediction are obtained by solving constrained non-negative least squares. To reduce the higher computational complexity per online round, two online learning (OL) approaches viz., weighted-follow-the-leader and weighted-follow-the-regularized-leader are proposed, inspired by the OP models. In OP, ASP difference (i.e, the gap between the ASP achieved by prediction and that by known content popularity) is bounded, while in OL, sub-linear MSE regret and linear ASP regret bounds are obtained. With MovieLens dataset, simulations verify that OP methods are better for MSE and ASP difference minimization, while the OL approaches perform well for the minimization of the MSE and ASP regrets.
引用
收藏
页码:1087 / 1100
页数:14
相关论文
共 29 条
[1]  
[Anonymous], 2014, Convex Optimiza- tion
[2]  
[Anonymous], 1995, SOLVING LEAST SQUARE
[3]   A low-complexity approach to distributed cooperative caching with geographic constraints [J].
Avrachenkov, Konstantin ;
Goseling, Jasper ;
Serbetci, Berksan .
Proceedings of the ACM on Measurement and Analysis of Computing Systems, 2017, 1 (01)
[4]   A Learning-Based Approach to Caching in Heterogenous Small Cell Networks [J].
Bharath, B. N. ;
Nagananda, K. G. ;
Poor, H. Vincent .
IEEE TRANSACTIONS ON COMMUNICATIONS, 2016, 64 (04) :1674-1686
[5]  
Blaszczyszyn B, 2015, IEEE ICC, P3358, DOI 10.1109/ICC.2015.7248843
[6]  
Bro R, 1997, J CHEMOMETR, V11, P393, DOI 10.1002/(SICI)1099-128X(199709/10)11:5<393::AID-CEM483>3.0.CO
[7]  
2-L
[8]   Grassmannian Differential Limited Feedback for Interference Alignment [J].
El Ayach, Omar ;
Heath, Robert W., Jr. .
IEEE TRANSACTIONS ON SIGNAL PROCESSING, 2012, 60 (12) :6481-6494
[9]   Popularity-Based Video Caching Techniques for Cache-Enabled Networks: A Survey [J].
Goian, Huda S. ;
Al-Jarrah, Omar Y. ;
Muhaidat, Sami ;
Al-Hammadi, Yousof ;
Yoo, Paul ;
Dianati, Mehrdad .
IEEE ACCESS, 2019, 7 :27699-27719
[10]   The MovieLens Datasets: History and Context [J].
Harper, F. Maxwell ;
Konstan, Joseph A. .
ACM TRANSACTIONS ON INTERACTIVE INTELLIGENT SYSTEMS, 2016, 5 (04)