Optimal Computation Offloading Decisions Based on System Utility and Cost Balance

Edge computing provides terminal users with computing resources and data processing capabilities by deploying edge nodes near Internet of Things (IoT) devices to meet the processing demands of terminal applications. In this study, we address the offloading problem of partitionable tasks in collaborative edge computing systems and propose a new evaluation model that jointly optimizes the time and energy costs along with the system utility. By calculating the full load rate of edge server clusters, we describe and evaluate the operational efficiency of the system, aiming to design a computation offloading decision scheme that not only reduces costs, but also enhances system utility. Based on the new evaluation model, an objective function is constructed to minimize the system’s cost-utility ratio, and an improved particle swarm optimization algorithm with enhanced search capabilities is proposed to solve it by designing a computation offloading strategy. Experiments conducted in a simulation based on the edge system of an industrial park verified that the cost-utility balanced optimization model achieves at least 7% reduction in the cost-utility ratio across varying user scales compared to traditional cost optimization models, while achieving over 20% reduction in optimization efficiency under different edge server scales. The computation offloading efficiency is significantly improved, effectively boosting the overall system performance through the collaborative optimization of resource efficiency and execution costs.