检索结果-内蒙古大学图书馆

A survey on design challenges of scheduling algorithms for wireless networks

INTERNATIONAL JOURNAL OF COMMUNICATION NETWORKS AND DISTRIBUTED SYSTEMS 2022年第3期28卷 219-265页

作者： Malekpourshahraki, Mojtaba Desiniotis, Christopher Radi, Marjan Dezfouli, Behnam Univ Illinois Dept Comp Sci BITS Networked Syst Lab Chicago IL 60680 USA Santa Clara Univ Dept Comp Engn Internet Things Res Lab Santa Clara CA 95053 USA Western Digital Res Milpitas CA USA

Recent advances in wireless technologies open up new avenues toward many newer applications;however, existing wireless networks are not efficient enough to satisfy some new applications' requirements such as low energy consumption, high throughput, and low end-to-end delay. scheduling algorithms are promising for performance improvement by addressing the limitations of standard networking protocols. Due to the widespread deployments and applications of wireless networks, a significant number of scheduling algorithms has been proposed to improve network performance. In this paper, we provide a systematic comparison among proposed scheduling algorithm with an exhaustive, network-independent framework to ease the comparison among different scheduler design and highlight important challenges in the term of algorithmic parameters, computational order, accuracy, and overhead. We also study a set of schedulers in two main categories: low-power, and high-throughput, concerning our framework to extract the weakness and strength of each proposed method compared to other scheduling algorithms.

关键词： scheduling algorithms low-power wireless networks high-throughput networks MAC

来源：评论

学校读者我要写书评

暂无评论

Usher: Utility-based scheduling algorithms for Polymorphic Applications 20

Usher: Utility-based Scheduling Algorithms for Polymorphic A...

引用

20th IEEE International Conference on Mobile Ad Hoc and Smart Systems, MASS 2023

作者： Wright, Kwame-Lante Bai, Fan Krishnamachari, Bhaskar University of Southern California United States General Motors Global R&d

ISBN: (纸本)9798350324334

There is a growing set of sensor-driven latency-sensitive applications that are exceeding the capabilities of what can be provided through cloud services and are driving the adoption of distributed edge computing. We present Usher, a framework for structuring and scheduling such applications to enable efficient utilization of computing resources across networked devices. Usher exploits the concept of polymorphic applications, which refers to varied implementations of an application to provide different tradeoffs. Under the Usher framework, we introduce two utility-based scheduling algorithms for polymorphic applications, namely the Simple Greedy (SG) and Bundle-based Greedy (BG) algorithms. These heuristic algorithms are based on techniques for list scheduling and submodular set function optimization, respectively. In this work, we jointly consider the job version selection and scheduling problems by modeling it as a multiple-choice, multi-dimension 0-1 knapsack problem with multiple knapsacks, which we show to be NP-hard. We demonstrate that the SG algorithm, despite its simplicity, performs well for the types of sensor-driven applications we are considering, even outperforming the BG algorithm in most cases. In fact, SG performs optimally for our YOLO-based object detection task dataset in all cases we evaluated with an exact solution. © 2023 IEEE.

关键词： scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

Rapid and Low-Cost Evaluation of Multi-fidelity scheduling algorithms for Hyperparameter Optimization 24th

Rapid and Low-Cost Evaluation of Multi-fidelity Scheduling...

引用

24th International Conference on Intelligent Data Engineering and Automated Learning, IDEAL 2023

作者： Mishra, Aakash Hsu, Jeremy D’Agostino, Frank Sirin, Utku Idreos, Stratos Harvard University CambridgeMA02138 United States

ISBN: (纸本)9783031482311

Hyperparameter optimization (HPO), the process of searching for optimal hyperparameter configurations for a model, is becoming increasingly important with the rise of automated machine learning. Using HPO can be incredibly expensive as models increase in size. Multi-fidelity HPO schedulers attempt to more efficiently utilize resources given to each model configuration by determining the order they are evaluated in and the number of epochs they are run for. Pre-tabulated benchmarks are often used to reduce the compute power required to evaluate state-of-the-art schedulers. However, over-reliance on evaluation using these benchmarks can lead to overfitting. To solve this problem, we introduce a Platform for Hyperparameter Optimization Search Simulation (PHOSS), that enables rapid HPO scheduler evaluation by dynamically generating surrogate benchmarks. PHOSS uses a computationally efficient and expressive exponential decay parametric modeling approach to accurately generate surrogate benchmarks from real-world dataset samples without having to train an expensive surrogate model. We demonstrate that PHOSS can simulate several state-of-the-art schedulers on real-world benchmarks 4.5 × faster while also dramatically reducing multi-GPU compute requirements by enabling full testing of HPO schedulers on a single commodity CPU. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.

关键词： scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

Risk-Aware scheduling algorithms for Variable Capacity Resources

Risk-Aware Scheduling Algorithms for Variable Capacity Resou...

引用

2023 International Conference on High Performance Computing, Network, Storage, and Analysis, SC Workshops 2023

作者： Perotin, Lucas Zhang, Chaojie Wijayawardana, Rajini Benoit, Anne Robert, Yves Chien, Andrew Laboratoire Lip Ens Lyon Lyon France Microsoft Research SeattleWA United States University of Chicago ChicagoIL United States University of Tennessee Knoxville United States Argonne National Laboratory United States

ISBN: (纸本)9798400707858

The drive to decarbonize the power grid to slow the pace of climate change has caused dramatic variation in the cost, availability, and carbon-intensity of power. This has begun to shape the planning and operation of datacenters. This paper focuses on the design of scheduling algorithms for independent jobs that are submitted to a platform whose resource capacity varies over time. Jobs are submitted online and assigned on a target machine by the scheduler, which is agnostic to the rate and amount of resource variation. The optimization objective is the goodput, defined as the fraction of time devoted to effective computations (re-execution does not count). We introduce several novel algorithms that: (i) decide which fraction of the resources can be used safely;(ii) maintain a risk index associated to each machine;and (iii) achieves a global load balance while mapping longer jobs to safer machines. We assess the performance of these algorithms using one set of actual workflow traces together with three sets of synthetic traces. The goodput achieved by our algorithms increases up to 10% compared to standard first-fit approaches, while we never experience any loss in complementary metrics such as the maximum or average stretch. © 2023 ACM.

关键词： scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

Performance Analysis of 5G NR Cell Using Classical scheduling algorithms and Different MIMO 8th

Performance Analysis of 5G NR Cell Using Classical Schedulin...

引用

8th International Conference on ICT for Sustainable Development, ICT4SD 2023

作者： Kumar, Rajesh Sinwar, Deepak Singh, Vijander Department of Computer and Communication Engineering Manipal University Jaipur Rajasthan Jaipur303007 India Department of ICT and Science Faculty of Information Technology and Electrical Engineering NTNU–Norwegian University of Science and Technology Ålesund Norway

ISBN: (纸本)9789819949311

5G NR has been deployed in most of the countries and covers three major use case scenarios, viz. enhanced mobile broadband (eMBB), ultra-reliable low latency communication (URLLC), and massive machine type of communication (mMTC). To enable the 5G services on the existing 4G infrastructure, the non-standalone architecture (NSA) of 5G NR is equally popular. For NSA deployment, many countries are utilizing the 3.5 GHz band on time division duplex (TDD) transmission. For both uplink and downlink transmissions, TDD is utilizing same frequency bands and scheduling of resources in such environments is quite challenging. Here, the performance of 5G NR is analysed on 20 MHz and 100 MHz channel bandwidth in terms of resource utilization by comparing the performance of different scheduling algorithms, i.e. Professional Fair, Round Robin, and Best CQI on different MIMO (64 × 64, 32 × 32, 16 × 16, 8 × 8) set-up and block error rates. The performance of Best CQI algorithm is found to be better than other scheduling algorithms under considerations. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd 2023.

关键词： scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

A Comprehensive Study of 5th Generation scheduling algorithms 3rd

A Comprehensive Study of 5th Generation Scheduling Algorithm...

引用

3rd International Workshop on New Approaches for Multidimensional Signal Processing, NAMSP 2022

作者： Kumar, Pardeep Singh, Parvinder Department of CSE DCRUST Murthal India

ISBN: (纸本)9789811978418

Now a days, the 5th-Generation communication standard for Standalone and Non-Stand-alone mode has been released in "Release 15" to "Release 18". This review paper is written to highlight all the key points of 5th generation communication technology covers core concepts of 5th generation communication network, its architecture for both Standalone and Non-Stand-alone mode, different existing scheduling algorithms used in 5G namely Proportional Fair (PF), Modified least weighted delay first (MLWDF), Exponential Proportional Fair (EX-PF), Frame Level Schedular (FLS), Round Robin (RR) and a detailed review of proposed scheduling algorithms anticipated by different researchers. This article is covering four different sections. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

Comparative Study of scheduling algorithms for Multiprocessor Systems

Comparative Study of Scheduling Algorithms for Multiprocesso...

引用

2023 IEEE International Conference on Intelligent and Innovative Technologies in Computing, Electrical and Electronics, ICIITCEE 2023

作者： Sreenath, M. Vijaya, P.A. Department of ECE BNMIT Bengaluru under VTU Belagavi India BNMIT Department of ECE Karnataka Bengaluru India

ISBN: (纸本)9781665462631

Traditional real-time scheduling algorithms are developed employing a standard framework to satisfy the requirements for a rapid execution time, high performance, and low power consumption. However, real-time control is gaining significance for a variety of applications, including manufacturing, data processing systems, aviation, and others. The goal of the current effort is to find the optimal functional analysis of the system while considering performance efficiency and processor use into account as critical components for tasks to be successfully executed within the allocated time. In this study, real-time scheduling methodologies and multiprocessor system scheduling algorithms are compared in order to determine the characteristics that are most useful for classification. The work load should be distributed equitably among the system's various processors in order to prevent faulty operations in an embedded system. Real-time scheduling algorithms are often built utilizing a standard framework in order to meet the demands for lowest execution time, minimal power consumption, and maximum performance. Users can implement new algorithms on various simulation tools. © 2023 IEEE.

关键词： scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

Benchmarking DAG scheduling algorithms on Scientific Workflow Instances 9th

Benchmarking DAG Scheduling Algorithms on Scientific Workfl...

引用

9th Russian Supercomputing Days International Conference, RuSCDays 2023

作者： Sukhoroslov, Oleg Gorokhovskii, Maksim Institute for Information Transmission Problems of the Russian Academy of Sciences Moscow Russia HSE University Moscow Russia

ISBN: (纸本)9783031494345

Many real-world applications executed on distributed computing systems are organized as directed acyclic graphs (DAGs) of tasks. The algorithm employed for scheduling these tasks across the system has a substantial impact on the achieved performance. Despite the numerous DAG scheduling algorithms proposed by researchers, there is a lack of benchmarks that evaluate the performance of such algorithms on a set of real application instances in realistic conditions. Thus developers of runtime systems often resort to the use of simple but inefficient algorithms. In this work we aim to fill this gap by proposing a benchmark for evaluating DAG scheduling algorithms based on a set of 150 real-world workflow instances with up to 1695 tasks and 10 realistic cluster configurations with multi-core machines. We apply this benchmark for evaluation of 16 scheduling algorithms including the well-known static algorithms and the commonly used in practice dynamic algorithms. The obtained results demonstrate that the proposed benchmark allows to clearly separate and compare the algorithms performance from different angles and to make important observations. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

Design and Performance Comparison of Latency Efficient scheduling algorithms for Smart Environment 4

Design and Performance Comparison of Latency Efficient Sched...

引用

4th International Conference on Communication, Computing and Industry 6.0, C216 2023

作者： Boomiga, S.S. Venkatesan, V. Prasanna Department of Artificial Intelligence and Data Science Sri Manakula Vinayagar Engineering College Puducherry India Department of Banking Technology Pondicherry University Puducherry India

ISBN: (纸本)9798350327328

Smart Environment aims to improve humans' quality of life. IoT is an emerging paradigm that makes the environment smart by using sensing technologies in IoT devices and cloud computing technology. Various surveys portray that IoT devices are growing rapidly which results in more data which leads to more delay in response time for the services provided in the smart environment since the Cloud server imposes more transmission latency. This transmission latency is reduced by the emergence of Edge computing in which the services can be provided by using Edge as a computing device. However, the Edge is a resource-constrained device, so the proposed work aims to identify the process that can be processed at the Edge based on context and priority based scheduling. Therefore, an algorithm is designed to execute the process that needs immediate attention by applying context and priority-based scheduling. In a comparative analysis of algorithms, the experimental results showcase that the proposed algorithms notably reduce latency in IoT processes. © 2023 IEEE.

关键词： scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

Resource Orchestration and scheduling algorithms for Enhancing Distributed Reinforcement Learning

Resource Orchestration and Scheduling Algorithms for Enhanci...

引用

2023 China Automation Congress, CAC 2023

作者： Li, Mingyang Wu, Fan Kou, Qian Qian, Long Chen, Xingyu Lan, Xuguang Xi' An Jiaotong University Institute of Artificial Intelligence and Robotics Shaanxi Xi'an China

ISBN: (纸本)9798350303759

In modern distributed computing, effective resource allocation assumes a paramount role in bolstering application performance. Kubernetes leverages localized scheduling algorithms to ensure astute resource utilization and alignment with application requisites. Nonetheless, the domain of distributed reinforcement learning poses challenges rooted in the intricate landscape of multi-node agent training and environment simulation, which engenders resource contention issues and heightened communication overhead. This paper introduces the LeaderFirst and Batch Tasks Centralized algorithms to counteract these challenges. The LeaderFirst algorithm demonstrates judicious allocation of Learner resources, disentangling reinforcement learning tasks to attain synchronization and proficiently alleviate competitive dynamics. Conversely, the Batch Tasks Centralized algorithm concentrates on the efficient bundling of tasks to maximize resource employment, all while judiciously addressing the mitigation of distributed communication overhead. Through empirical assessment, these algorithms showcase pronounced advantages over their Kubernetes native counterparts. © 2023 IEEE.

关键词： scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：