UCB Algorithm Based Decision Making Method for Edge Computing Tasks Offloading

As the demand for high-performance computing on mobile devices escalates, mobile edge computing has risen as a cutting-edge computing paradigm designed to curtail latency and enhance service quality. This paper introduces a decision-making approach for computational task offloading, leveraging the upper confidence bound algorithm to optimize energy consumption and execution latency in edge computing scenarios. To begin with, we construct an integrated system framework that includes a mobile device, a decision-making entity for offloading, an edge server, and a cloud server. Following this, we transform the offloading decision challenge into a multi-armed bandit scenario and implement the UCB algorithm to achieve a strategic equilibrium between exploring new possibilities and capitalizing on existing knowledge, consequently discovering the optimal offloading approach. Additionally, we introduce an objective function that concurrently minimizes delay and maximizes energy efficiency. The performance of our proposed algorithm is rigorously evaluated through simulation experiments, which demonstrate its superiority in reducing both the total system energy consumption and the latency associated with computational task offloading when compared to existing methodologies.