Optimized Deployment of Core Network Slices for Satellite-Terrestrial Integrated Networks

Abstract

The deployment of core network slices (NSs) in the satellite-terrestrial integrated network (STIN) is pivotal for enabling low-latency communications and catering to diverse service requirements. However, the limited satellite resources and the globally distributed users pose significant challenges for on-demand NSs deployment. In this letter, we model the network function deployment for multiple core NSs in the STIN as a mixed integer nonlinear programming problem (MINLP) with the objective of minimizing the average NS delay. We further propose a hybrid algorithm combining multi-swarm particle swarm optimization (MPSO) and genetic algorithm (GA) to address high dimensionality and variable coupling challenges in solving the optimization problem. The proposed algorithm utilizes the multi-swarm structure to jointly optimize multiple variables, where the strategy of each swarm is customized based on variable types for efficient optimization. Simulation results demonstrate the effectiveness of our algorithm.