Echo State Networks for Proactive Caching in Cloud-Based Radio Access Networks with Mobile Users

摘要

In this paper, the problem of proactive caching is studied for cloud radioaccess networks (CRANs). In the studied model, the baseband units (BBUs) canpredict the content request distribution and mobility pattern of each user,determine which content to cache at remote radio heads and BBUs. This problemis formulated as an optimization problem which jointly incorporates backhauland fronthaul loads and content caching. To solve this problem, an algorithmthat combines the machine learning framework of echo state networks withsublinear algorithms is proposed. Using echo state networks (ESNs), the BBUscan predict each user's content request distribution and mobility pattern whilehaving only limited information on the network's and user's state. In order topredict each user's periodic mobility pattern with minimal complexity, thememory capacity of the corresponding ESN is derived for a periodic input. Thismemory capacity is shown to be able to record the maximum amount of userinformation for the proposed ESN model. Then, a sublinear algorithm is proposedto determine which content to cache while using limited content requestdistribution samples. Simulation results using real data from Youku and theBeijing University of Posts and Telecommunications show that the proposedapproach yields significant gains, in terms of sum effective capacity, thatreach up to 27.8% and 30.7%, respectively, compared to random caching withclustering and random caching without clustering algorithm.