Energy-optimal resource scheduling and computation offloading in small cell networks

This paper provides a joint optimization framework of radio resource scheduling and computation offloading in small cell LTE based networks. We consider that mobile users are served by nearby small cell base stations which can be endowed with some computational capabilities. The objective is to minimize the average energy consumption at the user terminal to run its mobile applications, either locally or remotely, while satisfying average delay constraints tolerated by these applications. For this problem, we investigate offline dynamic programming approaches and we devise two solutions: deterministic and randomized, to find the optimal radio scheduling-offloading policy. We show that the dynamic offline strategies are able of achieving optimal energy efficiency at the mobile terminals. Indeed, they can adapt the processing decisions between: local processing, offloading, and staying idle, by exploiting their knowledge on the channel conditions and the application properties.