检索结果-内蒙古大学图书馆

IEEE Latin-American Conference on Communications, LATINCOM

作者： Carlos Fernandez Hernandez Oana Iova Fabrice Valois INSA Lyon Inria CITI UR3720 Villeurbanne France

ISBN: (数字)9798331521110

ISBN: (纸本)9798331521127

LoRaWAN is a key enabler for Internet of Things applications, providing long range communication at a low energy consumption. However, extensive use of downlink communication can reduce network capacity and increase uplink loss, as shown in several studies. The network server is responsible for scheduling downlink packets, while respecting strict timing constraints on the end devices and on the gateways. This paper presents the first evaluation of downlink scheduling implemented by two widely used LoRaWAN network servers: ChirpStack and The Things Stack. Using the ELoRa emulation tool, we inject real LoRaWAN traffic into these network servers and analyze their downlink scheduling algorithms based on defined performance metrics such as reliability and schedule efficiency. Our findings indicate that The Things Stack has a more accurate scheduler, but more conservative, as it delivers less downlink packets. ChirpStack, on the other hand, with its simpler algorithm, performs comparably and even equals The Things Stack as the number of end devices increases.

关键词： Performance evaluation Network servers Chirp scheduling algorithms Emulation LoRaWAN Downlink scheduling Timing Uplink

来源：评论

学校读者我要写书评

暂无评论

Mixed-Criticality Federated scheduling for Relaxed-Deadline DAG Tasks

Mixed-Criticality Federated Scheduling for Relaxed-Deadline ...

引用

Real-Time Systems Symposium (RTSS)

作者： Fei Guan Jinkyu Lee Chun Jason Xue Jen-Ming Wu Nan Guan Northeast Forestry University City University of Hong Kong Sungkyunkwan University Mohamed bin Zayed University of Artificial Intelligence Hon Hai Research Institute

ISBN: (数字)9798331540265

ISBN: (纸本)9798331540272

A mixed-criticality (MC) system is a computational platform shared by tasks with two or more safety-critical levels. An important research topic related to MC systems is designing scheduling algorithms that can satisfy the computation requirements of tasks with different criticality levels. Numerous studies have focused on this topic, but only a few have considered parallel tasks. To address the research gap, we propose a dual-criticality scheduling algorithm based on federated scheduling for parallel tasks with Directed Acyclic Graph (DAG) structures. We particularly focus on the task set in which each task has a deadline longer than its release period. To the best of our knowledge, our work is the first that does not assume the constrained-or implicit-deadline in the MC DAG task model. In addition to simulation experiments, we demonstrate that our algorithm has a capacity augmentation bound of 4, providing a quantitative worst-case performance guarantee for our algorithm.

关键词： Directed acyclic graph scheduling algorithms scheduling Real-time systems

来源：评论

学校读者我要写书评

暂无评论

QMLFQ: A Quantization-based Queue scheduling Framework for Efficient LLM Serving

QMLFQ: A Quantization-based Queue Scheduling Framework for E...

引用

Smart World Congress (SWC), IEEE

作者： Jiali Wu Hao Dai Jiashu Wu Wenming Jin Yang Wang Chinese Academy of Sciences Shenzhen Institute of Advanced Technology Shenzhen China University of Chinese Academy of Sciences Beijing China University of Southampton Southampton UK China Telecom Cloud Technology Co. Ltd Beijing China The SIAT-Suntang Big Data&AI Joint Innovation Laboratory Shenzhen China Shenzhen University of Advanced Technology Shenzhen China

ISBN: (数字)9798331520861

ISBN: (纸本)9798331520878

Fueled by advancements in Large Language Models (LLMs), applications like ChatGPT have ushered in a transformative era for AI-powered interactions. However, the computational demands of LLMs, characterized by their massive parameter size and intricate computations, pose a significant challenge to its model serving. Under high concurrency or resource-constrained scenarios, LLM service systems can experience substantial latency and potential crashes, jeopardizing user experience and system stability. To address this challenge, this paper presents a novel LLM serving framework that leverages dynamic queue scheduling and quantization switching to alleviate computational bottlenecks and reduce response latency. Specifically, we propose an enhanced scheduling algorithm based on the Multi-level Feedback Queue (MLFQ) that dynamically switches between various quantized-level models and prioritizes requests, thereby improving LLM inference parallelism and reducing job completion time (JCT) as well. Through extensive experiments conducted on open-source datasets, we demonstrate that our proposed approach effectively reduces JCTs compared to other leading baselines. This improvement in JCT ultimately leads to enhanced quality of service (QoS) for LLM-based applications.

关键词： Quantization (signal) scheduling algorithms Computational modeling Large language models Graphics processing units Quality of service Switches Dynamic scheduling User experience Stability analysis

来源：评论

学校读者我要写书评

暂无评论

Improving the Parameter Dependency for High-Multiplicity scheduling on Uniform Machines

arXiv

引用

arXiv 2024年

作者： Jansen, Klaus Kahler, Kai Pirotton, Lis Tutas, Malte Kiel University Department of Computer Science Germany

We address scheduling problems on uniform machines with high-multiplicity encoding, introducing a divide and conquer approach to assess the feasibility of a general Load Balancing Problem (LBP). Via reductions, our algorithm can also solve the more well-known problems QCmax (makespan minimization), QCmin (santa claus) and QCenvy (envy minimization). State-of-the-art algorithms for these problems, e.g. by Knop et al. (Math. Program.’23), have running times with parameter dependency pOmax(d2), where pmax is the largest processing time and d is the number of different processing times. We partially answer the question asked by Koutecký and Zink (ISAAC’20) about whether this quadratic dependency of d can be improved to a linear one: Under the natural assumption that the machines are similar in a way that smax/smin ≤ pOmax(1) and τ ≤ pOmax(1), our proposed algorithm achieves parameter dependency pOmax(d) for the problems Q{Cmax, Cmin, Cenvy}. Here, τ is the number of distinct machine speeds. Even without this assumption, our running times achieve a state-of-the-art parameter dependency and do so with an entirely different approach. © 2024, CC BY.

关键词： scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

A RESOURCE-CONSTRAINED STOCHASTIC scheduling ALGORITHM FOR HOMELESS STREET OUTREACH AND GLEANING EDIBLE FOOD

arXiv

引用

arXiv 2024年

作者： Artman, Conor M. Mate, Aditya Nwankwo, Ezinne Heching, Aliza Idé, Tsuyoshi Navrátil, Jiří Shanmugam, Karthikeyan Sun, Wei Varshney, Kush R. Goldkind, Lauri Kroch, Gidi Sawyer, Jaclyn Watson, Ian Machine Learning Group Lawrence Livermore National Laboratory LivermoreCA United States Microsoft New England CambridgeMA United States University of California Berkeley BerkeleyCA United States IBM Research - T.J. Watson Yorktown HeightsNY United States Fordham University BronxNY United States Leket Israel Ra'anana Israel Breaking Ground New YorkNY United States Change Machine BrooklynNY United States

We developed a common algorithmic solution addressing the problem of resource-constrained outreach encountered by social change organizations with different missions and operations: Breaking Ground-an organization that helps individuals experiencing homelessness in New York transition to permanent housing and Leket-the national food bank of Israel that rescues food from farms and elsewhere to feed the hungry. Specifically, we developed an estimation and optimization approach for partially-observed episodic restless bandits under k-step transitions. The results show that our Thompson sampling with Markov chain recovery (via Stein variational gradient descent) algorithm significantly outperforms baselines for the problems of both organizations. We carried out this work in a prospective manner with the express goal of devising a flexible-enough but also useful-enough solution that can help overcome a lack of sustainable impact in data science for social good. Copyright © 2024, The Authors. All rights reserved.

关键词： scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

Multi-Access Resource scheduling for Data and Relay Satellitess Based on a Genetic-Ant Colony Hybrid Algorithm

Multi-Access Resource Scheduling for Data and Relay Satellit...

引用

Satellite Internet (SAT-NET), International Conference on

作者： Ziyu Yao Jianmei Dai Jincheng Li Jiong Li School of Space Information Space Engineering University Beijing China

ISBN: (数字)9798350375176

ISBN: (纸本)9798350387469

The resource scheduling of Data and Relay Satellites plays a fundamental role in enhancing system performance, and increasing user satisfaction. This paper focuses on the scheduling of S-band multi-access link resources of Data and Relay Satellites. The objective is to maximize the completion of high-priority tasks and minimize the overall scheduling time. Drawing inspiration from the operational strategies of workshop scheduling models, a multi-objective scheduling model with time windows is constructed. Due to the high-dimensional mixed integer optimization nature of this problem, obtaining the global optimal solution is challenging. Therefore, this paper proposes a geneticant colony hybrid scheduling algorithm for multi-access resource scheduling. Simulation results demonstrate that, under the same task conditions, the genetic-ant colony hybrid algorithm improves scheduling performance by 8.78% and 5.38% compared to the single genetic algorithm and the single ant colony algorithm, respectively. Furthermore, it reduces scheduling processing time by 16.21% and 29.36%, respectively.

关键词： Satellites scheduling algorithms System performance Simulation scheduling Data models Relays Optimization Genetic algorithms Context modeling

来源：评论

学校读者我要写书评

暂无评论

Job scheduling in High Performance Computing Systems with Disaggregated Memory Resources

Job Scheduling in High Performance Computing Systems with Di...

引用

IEEE International Conference on Cluster Computing

作者： Jie Li George Michelogiannakis Samuel Maloney Brandon Cook Estela Suarez John Shalf Yong Chen Texas Tech University USA Lawrence Berkeley National Laboratory USA Jülich Supercomputing Centre Forschungszentrum Jülich Germany Institute of Computer Science University of Bonn Germany

ISBN: (数字)9798350358711

ISBN: (纸本)9798350358728

Disaggregated memory promises to meet growing memory requirements of applications while improving system resource utilization in high-performance computing (HPC) systems. Compared to traditional systems-where expensive resources such as CPUs, GPUs, and memory, are assigned to jobs in units of nodes-systems with disaggregated memory introduce memory pools that can be shared among jobs; this introduces new optimization metrics to the job scheduler. In this paper, we propose a data-driven approach to evaluate job scheduling and resource configuration in HPC systems with disaggregated memory. To incorporate the memory requirements of jobs for both local and disaggregated memory resources and improve system efficiency in open-science HPC systems, we introduce a novel job scheduling algorithm called FM (Fair Memory). Our simulation results show that FM outperforms commonly-used job schedulers in terms of jobs' bounded slowdown when the shared memory pool capacity is limited, and in terms of fairness under all conditions.

关键词： Degradation Production systems Frequency modulation Runtime scheduling algorithms High performance computing Simulation Memory management Resource management Optimization

来源：评论

学校读者我要写书评

暂无评论

An active distribution network optimization and scheduling method that takes into account fluctuations in new energy sources

An active distribution network optimization and scheduling m...

引用

Energy, Power and Electrical Technology (ICEPET), International Conference on

作者： Zhisong Xu Nanchuan Zhang Jie Li Chengjiao Zhong Kang Yao Jisheng Huang Rong Yuan Di Zhai Xiaoyun Mi Zhencun Li Yonghang Wang Rongyue Li Yuxi Power Supply Bureau Yunnan Power Grid Co. Ltd. Yuxi China LincangPower Supply Bureau Yunnan Power Grid Co. Ltd. Lincang China

ISBN: (数字)9798350352658

ISBN: (纸本)9798350352665

This article aims to explore and propose an active distribution network optimization and scheduling method that takes into account new energy fluctuations, in order to address the challenges that new energy fluctuations bring to traditional power systems. This method constructs an optimization model that takes into account the volatility characteristics of new energy sources such as wind and solar power through in-depth analysis, providing a mathematical basis for subsequent scheduling algorithms. By integrating advanced prediction technologies, precise prediction and dynamic scheduling of new energy output can be achieved to improve the stability and reliability of the distribution network. The experimental results indicate that the method proposed in this article is expected to provide new ideas and technical support for solving the challenges brought by the volatility of new energy, promote the intelligent and green development of active distribution networks, and make positive contributions to energy transformation and sustainable development.

关键词： Fluctuations scheduling algorithms Optimization models Green products Power system stability Dynamic scheduling Stability analysis

来源：评论

学校读者我要写书评

暂无评论

A reliability-aware and energy-aware task scheduling algorithm for heterogeneous multi-core systems

A reliability-aware and energy-aware task scheduling algorit...

引用

Chinese Control and Decision Conference, CCDC

作者： Dongyang Liu Junhua Chen Xueda Huang Haojun Hong School of Automation Chongqing University of Posts and Telecommunications Chongqing China

ISBN: (数字)9798350387780

ISBN: (纸本)9798350387797

Heterogeneous multi-core systems are widely used in high-performance computing applications. Reliability and low power consumption are important performance indicators of existing heterogeneous multi-core systems. The task scheduling algorithm with these two joint indicators as the optimization goal has important research significance. This paper designs a task scheduling algorithm for reliability and energy consumption awareness of heterogeneous multi-core systems. The algorithm is based on periodic real-time tasks with dependencies, and uses the redundancy of heterogeneous multi-cores to propose primary and secondary version backup technology for fault tolerance. The backup technology unifies the main task and the backup task and combines them with the DAG task model to construct a new task model. Reduce power consumption by introducing DVFS and DPM technology. DPM technology is mainly used for processor idle time, and consider using DVFS when the usage conditions of DPM are not met. In this paper, by comparing and evaluating randomly generated task models, the experimental results show that the algorithm has good performance in maintaining system reliability and reducing system power consumption.

关键词： Energy consumption Power demand scheduling algorithms Software algorithms Fault tolerant systems Software Software reliability

来源：评论

学校读者我要写书评

暂无评论

A Dynamic Resource scheduling Algorithm Based on Traffic Prediction for Coexistence of eMBB and Random Arrival URLLC

A Dynamic Resource Scheduling Algorithm Based on Traffic Pre...

引用

International Conference on Communication Technology (ICCT)

作者： Yizhou Jiang Xiujun Zhang Xiaofeng Zhong Shidong Zhou Department of Electronic Engineering Tsinghua University Beijing China Beijing National Research Center for Information Science and Technology State Key Laboratory of Space Network and Communications

ISBN: (数字)9798350363760

ISBN: (纸本)9798350363777

In this paper, we propose a joint design for the coexistence of enhanced mobile broadband (eMBB) and ultra-reliable and random low-latency communication (URLLC) with different transmission time intervals (TTI): an eMBB scheduler operating at the beginning of each eMBB TTI to decide the coding redundancy of eMBB code blocks, and a URLLC scheduler at the beginning of each mini-slot to perform immediate preemption to ensure that the randomly arriving URLLC traffic is allocated with enough radio resource and the eMBB traffic keeps acceptable one-shot transmission successful probability and *** framework for schedulers under hybrid-TTI is developed and a method to configure eMBB code block based on URLLC traffic arrival prediction is implemented. Simulations show that our work improves the throughput of eMBB traffic without sacrificing the reliablity while supporting randomly arriving URLLC traffic.

关键词： Time-frequency analysis Codes Upper bound scheduling algorithms Simulation Redundancy Stochastic processes Ultra reliable low latency communication Throughput Encoding

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：