Device-collaborative Computation Offloading in Mobile Edge Networks Based on Propagation Dynamics Theory

In the future Internet of Intelligent Things, when multiple devices have too much computation task at one moment, device-collaborative computation offloading helps them to reduce the task pressure while it may cause more users to suffer from task overload due to computation task delivery. Finding the critical conditions that lead to the all-user task overload, it helps us deploy the communication network to guarantee its stability. Therefore, in this paper we study the process of computation task delivery among the mobile edge networks (MENs) and its impact. We utilize the spreading threshold model in propagation dynamics theory to characterize the device-collaborative computation offloading. The edge-based compartmental theory is applied to describe the proposed model, establish the propagation dynamics equations, and derive the saddle points. Through analysis and simulation, we find that the network status that all users suffer from data overload can be avoided by controlling the proportion of users who choose device-collaborative computation offloading and the network heterogeneity.