High Efficient and Near-Optimal Binary Computation Offloading Strategy Based on Game Theory and Greedy Optimization

Mobile Edge Computing (MEC) is considered as a promising paradigm to overcome the computational constraints of mobile devices by offloading intensive tasks to nearby edge servers. As the number of users in the MEC network increases, users inevitably compete for limited wireless and computing resources, leading to a substantial escalation in the complexity of network resource allocation. Motivated by this challenge, this paper focuses on a multi-user binary computation offloading system in an MEC environment over quasi-static competitive wireless channels. We propose a non-cooperative game model, where user devices strategically optimize their offloading decisions to minimize total costs in terms of latency and energy consumption. Building upon this, the existence and feasibility of a Nash equilibrium is rigorously proved, thereby ensuring stability within the system. Furthermore, a distributed computation offloading algorithm is proposed based on game optimization, which enables user devices to adaptively attain balanced offloading strategies with minimal computational overhead. Extensive simulations validate the effectiveness of the proposed algorithm, demonstrating that it achieves near-optimal performance compared with the centralized optimization methods while avoiding additional server load or the need for user-specific configuration.