Distributed Optimization of Task Offloading and Resource Allocation for Mobile Edge Computing with Multifactorial Uncertainty

As the demand for computation-intensive and lowlatency services grows, mobile edge computing (MEC) has been widely applied in smart devices to provide efficient and real-time assistance. However, most existing studies impose fixed assumptions and lack consideration for the uncertainty within MEC. This makes it difficult for these studies to reasonably offload tasks in complex and highly volatile scenarios. Therefore, we construct an MEC task offloading system considering multifactorial uncertainty (MECTOS-MU), which involves multiple devices and MEC servers (MSs). In MECTOS-MU, task offloading and resource allocation are jointly optimized while complying with the constraint on latency to minimize the energy consumption of all devices, which is an NP-hard problem. To address this issue, we propose a novel algorithm called distributed game offloading based on load balancing (DGOLB). This method integrates task offloading prioritization, static game theory, and load balancing to formulate efficient task offloading decisions and resource allocation schemes. Extensive simulation results demonstrate that DGOLB outperforms other baseline algorithms in terms of energy consumption, ratio of dropped tasks, and average task response time, especially in scenarios with a large number of devices.