讲座题目：     Learning to Optimize Dynamic Content Caching

主 讲 人：     Dr. Gan Zheng, Loughborough University, UK

主办单位：     信息与电气工程学院

时    间：     2018年4月9日，14:00

地    点：     G202

主要内容：     

Caching popular contents at small cells has been proposed as a promising solution to reduce the backhaul demand and transmission delays, and thus alleviate many bottleneck problems in wireless content networks.  However, content popularity keeps changing over spatio-temporal domains. This calls for effective caching placement schemes that can adapt quickly to the dynamic content popularity. In this talk, we will discuss the joint optimization of content placement and BS densities to minimize the content missing probability in a multi-cluster heterogeneous network. We introduce the convex optimization method for the long-term offline cache placement, and the deep learning based approach for frequent online cache update. The numerical results demonstrate that the proposed algorithms can keep track of the variation of content popularity and maintain low missing probability.