星座
供应
计算机科学
卫星
选择(遗传算法)
卫星星座
GSM演进的增强数据速率
计算机网络
服务(商务)
电信
工程类
航空航天工程
人工智能
物理
天文
经济
经济
作者
Siru Chen,Lei Jiao,Konglin Zhu,Lin Zhang
标识
DOI:10.1109/tnse.2025.3578194
摘要
Facing the growing demand for Low Earth Orbit (LEO) edge services, in order to address the manageability and economic issues in LEO edge computing, this study introduces an innovative two-timescale optimization approach designed to dynamically optimize satellite access selection, user request dispatching, and service replica placement. Integrating both online and offline optimizations, our method adapts to real-time fluctuations in user demand and satellite resources, effectively managing long-term decisions such as service migration and replica placement. We formalize this optimization challenge as a finite-horizon, integer-variable problem, taking into account both switching costs and resource utilization. Through extensive experimentation, our approach is proven to significantly balance performance enhancement and resource efficiency, and we prove the approximation ratio for each time slot the competitive ratio for the long-term cost. Our work contributes to the understanding of multi-timescale optimization in LEO edge computing and pro-vides valuable insights for designing efficient control mechanisms in satellite-based systems.
科研通智能强力驱动
Strongly Powered by AbleSci AI