.png){kind=icon}
.png){kind=icon}
Return
An adaptive rank-based coevolutionary learning particle swarm optimization algorithm for server placement in edge computing
.png){kind=icon}
PRE
AI
DOI：10.1016/j.eswa.2026.132381.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Mobile edge computing (MEC) enables low-latency service provisioning by deploying resources at the network edge. In MEC, edge server placement (ESP) plays a pivotal role in optimally allocating heterogeneous servers to base stations (BSs). However, achieving an optimal ESP is hindered by significant challenges. In the modeling aspect, the accurate balance of cloud offloading and user latency is complicated by increasing service demands. From an algorithmic perspective, the development of a scalable and accurate algorithm is constrained by the exponential growth of decision variables in large-scale MEC networks. To address these issues, a constrained multi-objective model is formulated to jointly optimize delay, energy consumption, and load balancing. Furthermore, an adaptive rank-based coevolutionary learning algorithm is developed for ESP. Specifically, a domain-driven variable grouping (DVG) strategy is designed to effectively mitigate the curse of dimensionality. Subsequently, an adaptive rank-based coevolutionary learning (ARL) strategy is introduced to enhance solution precision in large-scale environments. Finally, experiments on the Shanghai Telecom and Australia dataset demonstrate that the proposed algorithm outperforms the presentative state-of-the-art algorithms.
Keywords:
Mobile edge computing
Edge server placement
Multi-objective optimization
Coevolutionary learning
Particle swarm optimization
Journal
IF:
7.5
Papers:
2.9W
Citations:
10.2W