Energy-Efficient Robust Resource Provisioning in Virtualized Wireless Networks

This paper proposes a robust resource allocation approach in virtualized wireless networks (VWNs) to address the uncertainty in channel state information (CSI) at the base station (BS) due to estimation error and mobility of users. In this set-up, the resources of an OFDMA-based wireless network are shared among different slices where the minimum reserved rate is considered as the quality-of-service (QoS) requirement of each slice. We formulate the robust resource allocation problem against the worst-case CSI uncertainty, aiming to maximize the overall energy efficiency (EE) of VWN in terms of a newly defined slice utility function. Uncertain CSI is modeled as the sum of its true estimated value and an error assumed to be bounded in a specific uncertainty region. The formulated problem suffers from two major issues: computational complexity and energy-efficiency degradation due to the considered error in the maximum extent. To deal with these issues, we consider a specific form of uncertainty region to solve the robust resource allocation problem via an iterative algorithm. The simulation results demonstrate the effectiveness of the proposed algorithms.