The paper presents a novel resource slicing scheme with cross-cell coordination in satellite-terrestrial integrated networks (STIN) to address the challenges posed by spatiotemporal dynamics in service demands and satellite mobility.
The key highlights are:
Formulation of the resource slicing problem as a long-term optimization problem to minimize the overall system cost in terms of resource usage and delay performance.
Proposal of a distributed resource slicing (DRS) scheme that decomposes the problem into two subproblems:
a) Resource reservation problem with a given satellite set in each slicing window.
b) Satellite selection problem for the slicing window in each cell.
Development of a hybrid data-model co-driven approach:
a) An asynchronous multi-agent reinforcement learning-based algorithm to determine the optimal satellite set serving each cell.
b) A distributed optimization-based algorithm to make the resource reservation decisions for each slice.
Simulation results demonstrate that the proposed DRS scheme outperforms benchmark methods in terms of resource usage and delay performance, with faster convergence speed compared to a fully data-driven approach.
다른 언어로
소스 콘텐츠 기반
arxiv.org
더 깊은 질문