Energy-Efficient Cooperation in Mobile Edge Computing-Enabled Cognitive Radio Networks

Abstract

Traditional mobile edge computing (MEC) methods always assume that the wireless devices (WDs) can offload their data to the base stations (BSs) or the access points (APs) at any time, which are not practical due to the tension between a large number of the WDs and the limited spectrum resources. In this paper, a framework for MEC-enabled cognitive radio (CR) networks is proposed, which integrates three technologies: MEC, CR, and wireless power transfer (WPT). To obtain the spectrum for offloading, cooperative relaying is considered. Optimization problems are formulated to study the upper bound of the energy efficiency (EE) of the WD and to maximize the practical EE in both partial offloading and local computing scenarios, which are non-convex and intractable. In order to tackle these problems, a two-phase method is proposed. The transmit power, the time for energy harvesting (EH) and MEC, and the central processing unit (CPU) frequency of the WD are jointly optimized. Semi-closed-form solutions are obtained in partial offloading scenario by using fractional programming theory, Lagrangian dual decomposition, and successive pseudo-convex approximation (SPCA) methods. Closed-form solutions are obtained for local computing scenarios. The simulation results show the effects of the different parameters on the system performance.