Research on Heterogeneous Edge Server Deployment Strategy Based on Traffic Prediction

Mobile edge computing (MEC), as an emerging computing paradigm, significantly enhances computational efficiency and reduces latency. In this paper, we address the heterogeneous server deployment problem in MEC. Leveraging historical traffic data from base stations, we employ the Prophet model to predict future network traffic. Based on these predictions, we propose a hybrid heuristic algorithm, PSA, which integrates Particle Swarm Optimization (PSO) and Simulated Annealing (SA). Experimental results demonstrate that our approach effectively reduces the average latency by 1–1.5 ms while achieving superior performance in terms of deployment cost and load balancing.