An efficient topology partitioning algorithm for system-level parallel simulation of mega satellite constellation communication networks

Article Computer Science, Information Systems

Energy-Efficient Multiprocessor-Based Computation and Communication Resource Allocation in Two-Tier Federated Learning Networks

Rukhsana Ruby et al.

Summary: In traditional federated learning (FL), edge devices jointly train a machine learning model by communicating learning updates without exchanging data samples. In this study, a two-tier FL network is considered, where IoT nodes are the core clients, low altitude aerial platforms (UAVs) serve as model aggregators at the middle tier, and a high-altitude aerial platform (UAV with relatively high altitude) acts as the top-most layer aggregator. The energy-efficient computation and communication resource allocation in this network is addressed using an iterative algorithm that solves separate subproblems and then jointly solves the entire problem. Offline and online client scheduling schemes are proposed for optimal edge node selection and workload assignment based on data quality and workload constraints. Extensive simulations with real data verify the effectiveness of the proposed resource allocation scheme and highlight the importance of considering both computation, communication, and model divergence weight in FL performance.

IEEE INTERNET OF THINGS JOURNAL (2023)

添加到收藏夹

Article Engineering, Electrical & Electronic

Performance Analysis of Satellite Communication System Under the Shadowed-Rician Fading: A Stochastic Geometry Approach

Dong-Hyun Jung et al.

Summary: This paper investigates downlink low Earth orbit (LEO) satellite communication systems, where multiple LEO satellites are uniformly distributed according to a homogeneous binomial point process (BPP). By analyzing distance distributions and serving satellite cases, the exact outage probability is derived for system throughput maximization under satellite-visibility and outage constraints.

IEEE TRANSACTIONS ON COMMUNICATIONS (2022)

添加到收藏夹

Article Computer Science, Theory & Methods

Accelerating Large-Scale Prioritized Graph Computations by Hotness Balanced Partition

Shufeng Gong et al.

Summary: Prioritized computation shows promising performance for a large class of graph algorithms by prioritizing important vertices. Hotness Balanced Partition (HBP) is proposed for accelerating prioritized iterative graph computations by distributing vertices based on their hotness to evenly distribute hot vertices among workers.

IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS (2021)

添加到收藏夹

Article Computer Science, Information Systems

Dynamic Graph Partitioning Scheme for Supporting Load Balancing in Distributed Graph Environments

Dojin Choi et al.

Summary: This paper proposes a novel incremental graph partitioning scheme based on vertex-cut, supporting load balancing and utilizing hot data for partitioning. The superiority of the proposed scheme is verified through performance evaluations, comparing it with existing partitioning schemes.

IEEE ACCESS (2021)

添加到收藏夹

Article Computer Science, Information Systems

Satellite Communications in the New Space Era: A Survey and Future Challenges

Oltjon Kodheli et al.

Summary: The survey captures the state of the art in SatComs and highlights the most promising open research topics. Main drivers of innovation, promising applications, in-depth literature review, future challenges and open research topics are presented.

IEEE COMMUNICATIONS SURVEYS AND TUTORIALS (2021)

添加到收藏夹

Article Engineering, Electrical & Electronic