An intelligent approach to partition multimedia traffic onto multiple radio access networks

Third Generation wireless networks and beyond will solicit the cooperation of heterogeneous access networks, in order to provide multimedia traffic to different classes of users, with varying quality requisites over regions and time zones. In this paper, the problem of how to efficiently partition the traffic demand onto the underlying radio access networks is addressed. The design objective is a resource allocation strategy, which provides a maximal resource utilization across all access networks, while at the same time respecting quality levels related to handover dropping performance, which can be predefined per service and per region. We propose a solution based on Reinforcement Learning, and report results. We extend the method to include the relative importance of each service, from the user's or the network providers' standpoint. This is done by making use of utility functions and maximizing the average aggregate utility.