Multi-Objective Resource Optimization for Hierarchical Mobile Edge Computing

Mobile edge computing (MEC) enables the computation of complex tasks efficiently on end-devices. It has been shown that MEC can potentially reduce the delay and offer energy-efficient operation. So far, the research has been focused on or limited to multi-server optimization for partial offloading or competing to offload to the best server among different sets of mobile device. In this paper, we propose a hierarchical structure where user devices that have available resources can act as servers without the need to offload to the cloud. Results show that this model improves the system utility. Furthermore, there is a tradeoff between the number of users in the network and the number of peers willing to accept tasks. As the number of users increase, the utility of the hierarchical system keeps increasing whereas the simple MEC system utility saturates.