检索结果-内蒙古大学图书馆

36th AAAI Conference on Artificial Intelligence / 34th Conference on Innovative Applications of Artificial Intelligence / 12th Symposium on Educational Advances in Artificial Intelligence

作者： Yuan, Mu Zhang, Lan Li, Xiang-Yang Univ Sci & Technol China Hefei Peoples R China

ISBN: (纸本)9781577358763

The cost efficiency of model inference is critical to real-world machine learning (ML) applications, especially for delay-sensitive tasks and resource-limited devices. A typical dilemma is: in order to provide complex intelligent services (e.g. smart city), we need inference results of multiple ML models, but the cost budget (e.g. GPU memory) is not enough to run all of them. In this work, we study underlying relationships among black-box ML models and propose a novel learning task: model linking. Model linking aims to bridge the knowledge of different black-box models by learning mappings (dubbed model links) between their output spaces. Based on model links, we developed a scheduling algorithm, named MLink. Through collaborative multi-model inference enabled by model links, MLink can improve the accuracy of obtained inference results under the cost budget. We evaluated MLink on a multi-modal dataset with seven different ML models and two real-world video analytics systems with six ML models and 3,264 hours of video. Experimental results show that our proposed model links can be effectively built among various black-box models. Under the budget of GPU memory, MLink can save 66.7% inference computations while preserving 94% inference accuracy, which outperforms multi-task learning, deep reinforcement learning-based scheduler and frame filtering baselines.

关键词： scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

Fairness scheduling for Tasks with Different Real-time Level on Heterogeneous Systems

Fairness Scheduling for Tasks with Different Real-time Level...

引用

International Conference on Parallel and Distributed Systems (ICPADS)

作者： Shifan Shao Shouzhen Gu Biao Sun Edwin H.-M. Sha Qingfeng Zhuge East China Normal University

For a real-time task-intensive systems, the fairness of task execution in dynamic scheduling is an important research area. However, many exist scheduling algorithms are unable to guarantee that tasks can be completed by the deadline and executed with a fair priority. In this paper, we proposed an efficient Multi-DAG real-time scheduling algorithm, HSDFW, which employs a fair priority calculation method to enable tasks with different real-time levels can be completed by the deadline, and a rejection policy to improve the performance of schedule. We proposed an INLP model and an evaluation simulator to verify the efficiency of HSDFW algorithm. The evaluation results show that our proposed algorithm has excellent performance in terms of average scheduling length and resource utilization.

关键词： Schedules scheduling algorithms Heuristic algorithms Dynamic scheduling Real-time systems Resource management Task analysis

来源：评论

学校读者我要写书评

暂无评论

Mixed-Criticality with Integer Multiple WCETs and Dropping Relations: New scheduling Challenges

Mixed-Criticality with Integer Multiple WCETs and Dropping R...

引用

Asia and South Pacific Design Automation Conference

作者： Federico Reghenzani William Fornaciari Politecnico di Milano Milano Italy

scheduling Mixed-Criticality (MC) workload is a challenging problem in real-time computing. Earliest Deadline First Virtual Deadline (EDF-VD) is one of the most famous scheduling algorithm with optimal speedup bound properties. However, when EDF-VD is used to schedule task sets using a model with additional or relaxed constraints, its scheduling properties change. Inspired by an application of MC to the scheduling of fault tolerant tasks, in this article, we propose two models for multiple criticality levels: the first is a specialization of the MC model, and the second is a generalization of it. We then show, via formal proofs and numerical simulations, that the former considerably improves the speedup bound of EDF-VD. Finally, we provide the proofs related to the optimality of the two models, identifying the need of new scheduling algorithms.

关键词： Fault tolerance Schedules Design automation scheduling algorithms Computational modeling Fault tolerant systems Numerical simulation

来源：评论

学校读者我要写书评

暂无评论

Polynomial-Size Models to Minimize Total Completion Time in a Parallel Batching Environment 10th

Polynomial-Size Models to Minimize Total Completion Time in ...

引用

10th IFAC Triennial Conference on Manufacturing Modelling, Management and Control (MIM)

作者： Druetto, Alessandro Grosso, Andrea Univ Torino Dipartimento Informat Via Pessinetto 12 I-10149 Turin Italy

We present a new integer linear formulation for the problem of minimizing the total completion time on a single parallel-batching machine. The new formulation is strong, in the sense that it delivers a sharp lower bound, and compact, i.e. polynomial in size, contrasted to recent successful models for the same problem that have exponential size and require to be handled by column generation. The new model is promising: combined with a rounding procedure, it allows to deliver good solutions with small, certified optimality gaps for instances with up to 50 jobs, and we believe it is susceptible of further improvements. Copyright (C) 2022 The Authors.

关键词： Operations Research scheduling algorithms Parallel Batching Total Completion Time Arc-Flow Models

来源：评论

学校读者我要写书评

暂无评论

Building Effective scheduling algorithms for Sensor Networks

Building Effective Scheduling Algorithms for Sensor Networks

引用

International Conference on Information Technology: New Generations (ITNG)

作者： Fangyang Shen Chunlei Liu Bing Qi Yanqing Ji Daniel Caban Department of CST NYC College of Tech Brooklyn NY Department of Math&CS Voldosta State University Voldosta GA Department of CS Methodist University Fayetteville NC Department of ECE Gonzaga University Spokane WA

ISBN: (纸本)9781479931897

In wireless sensor networks, some sensor nodes can switch to a sleep state to conserve energy if their neighbors can provide similar sensing coverage. In reality, when more sensor nodes sleep, a sensor network's coverage may degrade. In this paper, we evaluate the coverage-aware sleep scheduling algorithm in areas with uneven event distributions and study its performance on event detection rate and event detection delay. There are two major improvements in this paper: first, we add the components of event generator into this paper and collect data on event detection rate and event detection delay. Second, we analyze sleep probability and energy consumption for coverage-aware sleep scheduling algorithm. Compared with randomized deployment strategy, the density adaptive deployment performs better on event detection delay and event detection rate.

关键词： Event detection Delays scheduling algorithms Monitoring Educational institutions Energy consumption

来源：评论

学校读者我要写书评

暂无评论

Decision scheduling for Cloud Computing Tasks Relying on Solving Large Linear Systems of Equations

引用

COMPUTATIONAL INTELLIGENCE AND NEUROSCIENCE 2022年第1期2022卷 3411959页

作者： He, Jing Chongqing Ind Polytech Coll Coll Artificial Intelligence & Big Data Chongqing Peoples R China

With the continuous reform and innovation of Internet technology and the continuous development and progress of social economy, Big Data cloud computing technology is more and more widely used in people's work and life. Many parallel algorithms play a very important role in solving large linear equations in various applications. To this end, this article aims to propose and summarize a cloud computing task scheduling model that relies on the solution of large linear equations. The method of this paper is to study the technology of solving large-scale linear equations and propose an M-QoS-OCCSM scheduling model. The function of the experimental method is to solve the problem of efficiently executing N mutually dependent parallel tasks within limited resources, while fully satisfying users' expectations of task completion time, bandwidth rate, reliability, and cost. In this paper, the application experiment of large-scale linear equations in task scheduling is used to study task scheduling algorithms. The results show that when the task load is 10 and 20, the convergence speed of the MPQGA algorithm is 32 seconds and 95 seconds faster than that of the BGA algorithm, respectively.

关键词： scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

Various job scheduling algorithms in cloud computing: A survey

Various job scheduling algorithms in cloud computing: A surv...

引用

International Conference on Information Communication and Embedded Systems (ICICES)

作者： Yash P. Dave Avani S. Shelat Dhara S. Patel Rutvij H. Jhaveri Department of Computer Science and Engineering Shri Sa'd Vidya Mandal Institute of Technology Bharuch India Department of Computer Science and Information Technology Shri Sa'd Vidya Mandal Institute of Technology Bharuch India

ISBN: (纸本)9781479936984

Cloud Computing provides a Computing environment where different resources, infrastructures, development platforms and software are delivered as a service to customers virtually on pay per time basis. Low cost, scalability, reliability, utility-based computing are important aspects of cloud computing. Job scheduling is an essential and most important part in any cloud environment. With increasing number of users, Job scheduling becomes a strenuous task. Ordering the jobs by scheduler while maintaining the balance between quality of services (QoS), efficiency and fairness of jobs is quite challenging. scheduling algorithms are implemented considering parameters such as throughput, resource utilization, latency, cost, priority, computational time, physical distances, performance, bandwidth, resource availability. Though there are different scheduling algorithms available in cloud computing, a very less comparative study has been done on performance of various scheduling algorithms with respect to above mentioned parameters. This paper aims at a comparative study of various types of job scheduling algorithms that provide efficient cloud services.

关键词： Cloud computing scheduling algorithms scheduling Algorithm design and analysis Classification algorithms Resource management

来源：评论

学校读者我要写书评

暂无评论

Profit Maximization in Mobile Crowdsourcing: A Competitive Analysis

引用

IEEE ACCESS 2021年 9卷 27827-27839页

作者： Li, Qi Cai, Lijun Xu, Huanle Meng, Tao Hunan Univ Coll Comp Sci & Elect Engn Changsha 410082 Peoples R China Chinese Univ Hong Kong Dept Informat Engn Hong Kong Peoples R China Cent South Univ Forestry & Technol Coll Comp & Informat Engn Changsha 410004 Peoples R China

Mobile Crowdsourcing, which uses a crowd of workers to complete computer-complexity tasks, has become more and more popular. How to maximize the total profit, which is the difference between total task utility and total worker costs, is a key design goal in such crowdsourcing system. However, this problem is extremely hard because even in the offline situation, this problem has been proved to be NP-hard. In the online situation, we need to consider not only future unknown task arrivals, but also the heterogeneity of both tasks and workers. In this paper, we address this challenging problem from following aspects. To solve this problem, we first build an optimization framework to formulate this problem by explicitly modelling task utility and worker costs. Then, under this framework, we design efficient online assignment algorithms to assign tasks to workers. To analyze the efficiency of our proposed algorithms, we develop a dual-fitting method to analyze both primal objective and dual objective. We prove that our designed algorithms can achieve a constant competitive ratio of 2. Finally, we conduct extensive trace-driven simulations to demonstrate that the online algorithms can outperform baseline algorithms by nearly 20% in terms of the total profit achieved.

关键词： Task analysis Crowdsourcing Cost function Analytical models scheduling algorithms Programming scheduling Task utility worker costs online task assignments dual-fitting method competitive performance

来源：评论

学校读者我要写书评

暂无评论

Distributed Intelligent scheduling Algorithm for Wireless Communication Network Link Resources Based on Data Analysis

Distributed Intelligent Scheduling Algorithm for Wireless Co...

引用

Integrated Circuits and Communication Systems (ICICACS), IEEE International Conference on

作者： Yangjie Feng Information Science Research Institute of China Electronics Technology Corporation Beijing China

The construction of wireless network is the mainstream direction of future communication development, and link optimization provides a basis for information exchange between the next generation of intelligent mobile terminals and base stations. This paper studies the link scheduling problem based on data analysis. First, after deriving the traditional optimal power allocation strategy and improving the genetic algorithm to obtain a new model, this paper proposes an optimization scheme calculation that is suitable for resource sharing among nodes, and can also consider the impact of time utility and other factors on the global availability, network topology and quality of service to minimize the overhead, and applies it to the wireless communication network for example verification. The experimental results show that the distributed intelligent scheduling algorithm for wireless communication network link resources based on data analysis performs very well in terms of running time and system processing efficiency, which can reach more than 90% efficiency.

关键词： Performance evaluation Data analysis scheduling algorithms Wireless networks Quality of service scheduling Routing protocols

来源：评论

学校读者我要写书评

暂无评论

Data Cloud Platform of Industrial Internet of Things based on TSN scheduling Algorithm

Data Cloud Platform of Industrial Internet of Things based o...

引用

Integrated Circuits and Communication Systems (ICICACS), IEEE International Conference on

作者： Guangfu Zhang Jiajing Cai Xiuyan Zheng College of Information Engineering Hainan Vocational University of Science and Technology Haikou China

With the rapid development of Industrial Science and technology, industrial scientific data has the characteristics of huge scale, wide range, diverse types and strong timeliness. The Industrial Internet of things (IIOT) data cloud platform came into being. Based on this, this paper studies and analyzes the application of TSN scheduling algorithm (SA) in the IIOT data cloud platform. Firstly, the architecture and key technologies of the Internet of things are introduced; Then the SA and cloud computing (CC) tasks in CC are analyzed, and the workflow system architecture under the data cloud platform of the IIOT and the workflow system architecture under the CC environment are discussed; The TSN SA is proposed and applied to the data cloud platform of the IIOT; Finally, the reliability of TSN SA is evaluated and tested. The test results show that the TSN resource SA proposed in this paper can basically achieve the packet loss rate of less than 1%, that is, the successful packet reception rate can basically reach more than 99%. It is verified that the TSN SA proposed in this paper can ensure the reliability and real-time of data transmission. The results show that the synchronization effect can fully meet the needs of industrial applications.

关键词： Cloud computing Job shop scheduling scheduling algorithms Systems architecture Real-time systems Reliability Synchronization

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：