A Task Allocation Method in Edge Computing Based on Multi-Objective Optimization

Abstract

Edge computing has been widely used in many scenarios because it is able to improve the performance of cloud computing. With the feature of distribution in edge computing, computing tasks can be allocated to edge computing system for reducing the workload of cloud centers. And to compute these tasks in edge nodes can make some aggregation and calculation process happen close to users. Therefore, some transmission time can be saved when data from edge nodes are sent to users compared with that when data are sent from cloud centers. And it is important to find appropriate task allocation strategies because of less computing resources in edge devices. In this work, a task allocation method AWHA is put up for allocating tasks to edge nodes. It optimizes total time cost, computing resource and storage utilization for computing these tasks in edge nodes. Then, for the scenario where the computation results of each edge node need to be aggregated and do further calculation, a result aggregation strategy FDFA is put up for optimizing the aggregation process of results by allocating all the calculation process to each edge computing node. The experiment results show that AWHA and FDFA methods have optimizing ability.