Energy-saving Offloading Strategies for Mobile Cloud Tasks Based on Three-way Game Theory in Wireless Communications
Chunmao Jiang and Yue Liu
Mobile networks (4G/5G) cover almost everywhere, and they carry a tremendous amount of computation. As their resources are still limited, they are still struggling to meet the demands of compute-intensive tasks. Offloading mobile devices’ tasks using wireless networks to the cloud (edge cloud/far cloud) becomes an essential means to improve mobile devices’ performance. By analyzing the mobile cloud task offloading process, we found numerous three granular phenomena. A collection of mobile tasks on each device can perform its tasks locally, at the edge node, and in the far cloud, respectively. There is mutual gaming between the devices due to shared bandwidth and other factors. This paper proposes an offloading model for mobile cloud tasks, drawing on three-way game theory ideas. First, we define the cost of offloading mobile devices. Second, through a three-way game, the model divides the mobile cloud tasks into different locations rationally and efficiently. The game determines the places where tasks are offloaded, i.e., the set of tasks is divided into three parts by a pair of thresholds, which are offloaded to three locations, i.e., local, edge nodes, and cloud servers. Different devices will have different thresholds for partitioning, and we repeat iterations of adjusting both thresholds to find the optimal range of threshold partitioning based on the cost function. The simulation experiments results show that the strategy and the corresponding algorithm are beneficial and can optimize the energy consumption well under the premise of meeting the quality of service.
Keywords: Offloading strategy; Energy consumption; Three-way decisions; Game theory