Optimizing Resource Allocation in Edge Computing to Reduce Delay Based on Multi-Layer Particle Swarm Optimization

With the popularity of Internet of Things (IoT) applications and the explosion of mobile devices, the computation demands of data have increased. All compute requests are sent to a remote cloud server for processing, leading to more delay. The emergence of edge computing (EC) can store and compute data generated by network edge devices directly on that device. Nevertheless, implementing compute-intensive tasks on edge devices with limited storage and compute resources makes task allocation more challenging. The complexity of resource allocation can lead to problems such as system performance degradation and time delay. This paper proposes a resource allocation strategy optimized by multi-layer particle swarm optimization (MLPSO) for edge computing to reduce delay. First, we adopt an EC scenario where multiple mobile users are allocated to multiple edge servers dynamically. Then, to construct the delay minimization model, a non-orthogonal multiple access (NOMA) protocol and resource allocation strategy are adopted, which considers the interference between channels. Finally, the delay minimization model is optimized by using MLPSO, that have great potential to help particle swarms jump out of local optima, enhancing the ability to obtain optimal resource allocation strategy for EC. The simulation results, along with comparisons to other methods, clearly demonstrate the effectiveness of the proposed approach in significantly reducing the delay of the EC model.