Prioritized computation offloading and resource optimization for networks with strict latency

Abstract

Mobile Edge Computing (MEC) has improved capabilities of mobile devices through computation offloading. However, this has also increased latency from overheads in moving the computation around. While there is a plethora of strategies for binary offloading decisions, in this work we take a different approach by partially offloading computations based on priority. We divide the problem into offloading selection, transmission power optimization and spectrum and computing resource allocation and solve them sequentially. Our results show that this mechanism outperforms existing offloading schemes.