Content-Aware Multi-Subtask Offloading: A Coalition Formation Game-Theoretic Approach

Content-Aware Multi-Subtask Offloading: A Coalition Formation Game-Theoretic Approach

Abstract:

In existing work focusing on task offloading, the internal relationship between the offloaded task and its content is ignored but significant. Each subtask which is divided according to the type of application and content, obtains a corresponding delay requirement decided by the unique content characteristics of subtasks. Due to the limited computing capacity of mobile devices, it causes different delay of each subtask. In this letter, we first study a content-aware multi-subtask offloading problem based on the individual characteristics of each subtask. Under the different delay requirements of subtasks, the goal is to minimize the aggregate energy consumption by optimizing the offloading decision and channel allocation. The problem is formulated as a coalition-based subtask cooperation offloading game, which is proved with a stable coalition structure under our proposed cooperation order. Then, the content-aware task offloading (CATO) algorithm is proposed to obtain a stable coalition partition. Finally, simulations demonstrate the validity and effectiveness of the coalition-based subtask cooperation offloading game as well as the CATO algorithm.