TriStack: Efficient Stackelberg Game-Based Offloading for Cloud-Edge-Terminal Computing

Cloud-edge-terminal computing has become a promising approach to improve resource utilization and service for communication and network applications. It is crucial to rationalize task offloading and resource allocation to better satisfy quality of experience (QoE). However, without strong incentives, cloud servers (CSs) and edge servers (ESs) may be reluctant to assist user terminals (UTs). In this paper, we propose TriStack, a three-stage Stackelberg game-based offloading incentive mecha-nism for the cooperation and competition among the cloud server, edges servers and user terminals. We optimize the benefits of game players under the comprehensive consideration of economic factors, delay sensitivity and energy consumption cost, and prove the existence of Nash equilibrium (NE) at each stage using backward induction. In addition, TriStack is designed to find Stackelberg equilibrium (SE) that maximizes the benefits of CS, ESs, and UTs. The simulation results show that TriStack has good convergence. Compared with uniform pricing strategy and two-stage game, TriStack can enhance players' benefits by 11.45% and 20.07 %, respectively.