Leveraging Heterogeneous Wireless Big Data for Cross-layer  Resource Optimization in Cellular Networks

Harnessing different types of big data in different timescales from cellular networks to boost the performance of mobile networks remarkably is gaining wide attention from academia,mobile operators, and vendors, which is expected to provide new freedoms in system design from various aspects. One promising way to leverage the big data from different layers in wireless networks is to increase spectrum and energy resource utilization efficiency by predicting user behavior. This paper provides a survey of the research along this line. We first introduce the predictable information capturing collective and individual user behavior that are essential for managing caching and transmission resources in wireless edge. We then address two kinds of representative techniques using the predicted behavior information, namely proactive wireless edge-caching and predictive radio resource allocation. Finally, we identify open problems and future research topics.