Online Control of Traffic Split and Distributed Cell Group State Decisions for Multi-connectivity

Abstract

This paper considers the problem of joint control of traffic split and cell group state decision for downlink communication via multi-connectivity in a cellular network. In this problem, a master node makes a decision on traffic split every T slots, and each secondary base station determines its cell group state between 'activated' and 'deactivated' to save UE power consumption. To ensure independent control of each cell group through its hosting base station, cell group state decision should be independent of each other, evolving asynchronously cell group state across cell groups. State decision for each cell group is performed dynamically according to changing environments. Such aperiodic state decision and asynchronous state evolution cause considerable difficulty in developing algorithms to solve the problem. To overcome the difficulties, we employ a frame-based Lyapunov optimization framework with variable frame sizes and develop online algorithms that are simple to implement but provably provide near optimal values. Simulation results are shown that our algorithm outperforms the other competitive schemes.