检索结果-内蒙古大学图书馆

Bioinspired Adaptive Resource scheduling for QoS in Mobile Edge Deployments

IET Communications 2025年第1期19卷

作者： Kaur, Gagandeep Singh, Balraj Faheem, Muhammad School of Computer Science & Engineering Lovely Professional University Punjab Phagwara India School of Technology and Innovations University of Vaasa Vaasa Finland VTT Technical Research Center of Finland Ltd. Espoo Finland

As mobile edge computing (MEC) expands, efficient resource allocation and job scheduling become increasingly important. Existing techniques are frequently unable to offer acceptable quality of service (QoS), owing to inflexible scheduling algorithms and insufficient consideration of complex task and resource metrics. To overcome these constraints, this work proposes a novel adaptive vector autoregressive moving average with exogenous variables (VARMAx)-based bioinspired resource scheduling model designed specifically for mobile edge deployment. The proposed approach applies the resilient concepts of flower pollination optimisation (FPO) to map tasks to virtual machines (VMs), a technique that is sensitive to a wide variety of task variables such as makespan, deadline and CPU needs. Simultaneously, VM characteristics such as million instructions per second (MIPS), amount of cores, random access memory (RAM), availability and bandwidth are all taken into account, resulting in a more nuanced and adaptive scheduling process. Furthermore, a VARMAx model is included for task pre-emption, which assists in the recalibration of future VM capabilities, hence improving overall scheduling efficiency, particularly in real-time deployments. The suggested model outperforms existing techniques. Our results show an 8.3% reduction in makespan, a 4.5% improvement in deadline hit ratio, an 8.5% increase in energy efficiency, and a 10.4% increase in throughput. The huge improvements highlight the model's adaptability and efficacy, resulting in important advances in the field of QoS-aware task scheduling for MEC. This work represents a significant advancement in the field of effective resource scheduling, with the potential to guide future research and development efforts in mobile edge deployments. © 2025 The Author(s). IET Communications published by John Wiley & Sons Ltd on behalf of The Institution of Engineering and Technology.

关键词： scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

Intelligent Student Management Information System Based on Genetic Algorithm 3

Intelligent Student Management Information System Based on G...

引用

3rd International Conference on Artificial Intelligence and Computer Information Technology, AICIT 2024

作者： Liu, Yunshan Wuxi Vocational Institute of Commerce Jiangsu Province Wuxi China

ISBN: (纸本)9798350365443

In order to solve the problem that existing scheduling software cannot generate the best scheduling plan under multiple constraints as schools continue to expand, the author proposes an intelligent student management information system based on genetic algorithm. Using genetic algorithm as the core algorithm for scheduling processing, studying various constraint conditions, constructing fitness functions, realizing automatic generation of scheduling plans, and using this as the center, constructing an intelligent scheduling system based on the network to solve the scheduling problem of large university academic affairs departments. The test results show that under different scheduling conditions and rules, the satisfaction rate of student course selection reaches 98%, and the average overall rule satisfaction rate reaches 95.8%. The overall satisfaction rate of course scheduling efficiency in the intelligent scheduling system has been improved to 98%. Conclusion: This method provides new ideas and methods for intelligent course scheduling problems, making the management system of university course scheduling more intelligent, improving the rational allocation of teaching resources and the completion of educational and teaching plans. © 2024 IEEE.

关键词： scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

Three-Level Planning and Dispatching Method of Regional Comprehensive Energy Based on Stochastic Forest Algorithm

Three-Level Planning and Dispatching Method of Regional Comp...

引用

2024 International Conference on Artificial Intelligence and Power Systems, AIPS 2024

作者： Wang, Kai Zhang, Huailiang Dong, Weiwei Shanghai Saga Energy Technology Co. Ltd. Shanghai200240 China Shandong Xinfa Group Co. Ltd. Shandong Liaocheng252100 China School of Economics and Management Shanghai Institute of Technology Shanghai201418 China

ISBN: (纸本)9798350392210

At present, the three-level planning and dispatching mechanism of regional comprehensive energy are mostly set as a one-way structure, and the actual dispatching range is small, which leads to an increase in the final regional energy consumption. Therefore, the design and verification analysis of the three-level planning and dispatching method of regional comprehensive energy based on the random forest algorithm is proposed. According to the current testing requirements and changes in standards, we define the basic energy planning and scheduling indicators, adopt a multi-objective approach, expand the actual scheduling scope, and establish a multi-objective reverse planning and scheduling mechanism. On this basis, the planning and scheduling model of regional comprehensive energy three-layer random forest measurement is designed, and the planning and scheduling are finally realized by adaptive processing. The test results show that for the six selected test areas, the final energy consumption is controlled at 350 m3, The following shows that with the help of random forest calculation, the three-layer planning and dispatching method of regional comprehensive energy designed at present is more efficient, specific, and targeted with practical application value. © 2024 IEEE.

关键词： scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

Deep Reinforcement Learning Approach for Fairness-aware scheduling in Wireless Networks 13

Deep Reinforcement Learning Approach for Fairness-aware Sche...

引用

13th International Conference on Information and Communication Technology Convergence, ICTC 2022

作者： Kim, Minseok Hwang, Sangwon Lee, Inkyu School of Electrical Engineering Korea University Seoul Korea Republic of Central Technology Appraisal Institute Korea Technology Finance Corporation Seoul Korea Republic of

ISBN: (数字)9781665499392

ISBN: (纸本)9781665499392

This paper studies a scheduling problem in wireless network systems which have combinatorial and time-varying properties. Although studies on efficient scheduling algorithms have been widely investigated, a huge trade-off between the performance and complexity still exists. Furthermore, it is becoming necessary to consider quality of service (QoS) and fairness constraints as growing demands on wireless networks. To this end, this paper propose a deep reinforcement learning (DRL) approach which can maximize system throughput while considering QoS and fairness by designing DRL structure according to the objective function and constraints. Numerical results demonstrate the effectiveness of the proposed approach. © 2022 IEEE.

关键词： scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

CHS-Scheduler: A Hybrid Scheduler for QoS Task scheduling in Cloud Human Service Scenarios 7

CHS-Scheduler: A Hybrid Scheduler for QoS Task Scheduling in...

引用

7th International Conference on Cloud Computing and Big Data Analytics, ICCCBDA 2022

作者： Zhang, Renyu Zhu, Zongwei University of Science and Technology of China School of Software Engineering Suzhou China

ISBN: (纸本)9781665487115

With the rapid advancement of crowdsourcing and services technology, human service ecosystem is evolving rapidly and scenarios within it are diversifying. Represented by consulting service platforms on the cloud, task scheduling in such cloud human service scenarios has more key features than traditional QoS cloud scheduling scenarios, such as dynamic human processing capability and statistical property in task flows. Rule based scheduling heuristics usually produces poor scheduling results in multi-Target scenarios and have limited adaptability as using local scheduling information. scheduling algorithms based on meta-heuristics is time-consuming and tend to fall into local optima prematurely when facing large-scale task flows. More importantly, these scheduling methods pay little attention to optimizing those key features mentioned above. Thus, an efficient task scheduling approach is urgently needed as a key to future computing systems for these scenarios. In this paper, we propose CHS-Scheduler, a hybrid scheduler for QoS task scheduling in cloud human service scenarios. Firstly, a scheduling model focusing on statistical property in task flows and dynamic human processing capability is constructed. Secondly, a two-stage scheduling methodology is given. Stage one is a task scheduling algorithm based on static & dynamic hybrid dimensionality reduction mechanism. Stage two is a solution domain mapping mechanism based on incremental learning. Experiments show that CHS-Scheduler exhibits state-of-The-Art performance in both batch scheduling and online scheduling. © 2022 IEEE.

关键词： Dimensionality reduction Crowdsourcing Cloud computing scheduling algorithms Heuristic algorithms Ecosystems Focusing

来源：评论

学校读者我要写书评

暂无评论

Survey on Task scheduling Methods in Cloud RPS System 23

Survey on Task Scheduling Methods in Cloud RPS System

引用

23rd International Seminar on Intelligent Technology and Its Applications, ISITIA 2022

作者： Ciptaningtyas, Henning Titi Shiddiqi, Ary Mazharuddin Purwitasari, Diana Institut Teknologi Sepuluh Nopember Information Technology Department Surabaya Indonesia Institut Teknologi Sepuluh Nopember Informatics Department Surabaya Indonesia

ISBN: (数字)9781665460811

ISBN: (纸本)9781665460811

The cloud platform is a popular technology because of its prominent feature, scalability, which allows users to easily adjust the usage of cloud resources. Cloud Service Provider (CSP) uses cloud provisioning methods to provide a sufficient amount of virtual resources the user needs on time, then uses a task scheduling algorithm to process the workload within the deadline or budget constraint. Ineffective scheduling may lead to underutilization (which wastage of cloud resources) or overutilization (which degrades cloud performance). This paper presents the optimization metric and comparison of task scheduling algorithms in the Cloud Resource Provisioning with scheduling (Cloud RPS) system. We recommend using the hybrid algorithm to achieve the desired objective based on the algorithm that has been compared in this paper. © 2022 IEEE.

关键词： scheduling algorithms

来源：评论

学校读者我要写书评

暂无评论

Deep Reinforcement Learning for System-on-Chip: Myths and Realities

引用

IEEE ACCESS 2022年 10卷 98048-98064页

作者： Sung, Tegg Taekyong Ryu, Bo EpiSys Sci Inc Poway CA 92064 USA

Neural schedulers based on deep reinforcement learning (DRL) have shown considerable potential for solving real-world resource allocation problems, as they have demonstrated significant performance gain in the domain of cluster computing. In this paper, we investigate the feasibility of neural schedulers for the domain of System-on-Chip (SoC) resource allocation through extensive experiments and comparison with non-neural, heuristic schedulers. The key finding is three-fold. First, neural schedulers designed for cluster computing domain do not work well for SoC due to i) heterogeneity of SoC computing resources and ii) variable action set caused by randomness in incoming jobs. Second, our novel neural scheduler technique, Eclectic Interaction Matching (EIM), overcomes the above challenges, thus significantly improving the existing neural schedulers. Specifically, we rationalize the underlying reasons behind the performance gain by the EIM-based neural scheduler. Third, we discover that the ratio of the average processing elements (PE) switching delay and the average PE computation time significantly impacts the performance of neural SoC schedulers even with EIM. Consequently, future neural SoC scheduler design must consider this metric as well as its implementation overhead for practical utility.

关键词： Reinforcement learning Minimization Schedules Resource management Performance gain Clustering algorithms scheduling algorithms Deep learning Deep reinforcement learning heuristic scheduler neural scheduler resource allocation system-on-chip scheduling

来源：评论

学校读者我要写书评

暂无评论

Hybrid NOMA in Multi-Cell Networks: From a Centralized Analysis to Practical Schemes

引用

IEEE-ACM TRANSACTIONS ON NETWORKING 2022年第3期30卷 1268-1282页

作者： Hussein, Abdalla Rosenberg, Catherine Mitran, Patrick Univ Waterloo Dept Elect & Comp Engn Waterloo ON N2L 3G1 Canada

We investigate the performance of a hybrid non-orthogonal multiple access (NOMA) multi-cell downlink system (called hybrid as different users can have different successive interference cancellation (SIC) capabilities) by first formulating and solving a centralized proportional fair scheduling genie-assisted problem that jointly performs user selection, power allocation, power distribution, and modulation and coding scheme (MCS) selection. While such a genie is practically infeasible, it upper bounds the achievable performance. The results indicate that hybrid NOMA with a maximum of 2 multiplexed users can bring significant gains over a traditional OMA system (as long as enough users have the maximum SIC capability). Additionally, results show that the simple equal power allocation scheme (often used in the literature) yields performance lower than half the upper bound. Thus, we propose a simple static coordinated power allocation scheme across all cells for NOMA using a simple power map that is easily calibrated offline and show that with the calibrated power map, performance improves by 80%. Finally, we focus on the online scenario and propose a family of practical scheduling algorithms, each of them exhibiting a different trade-off between complexity (i.e., run-time) and performance.

关键词： NOMA Resource management Silicon carbide Power distribution Intercell interference Signal to noise ratio scheduling algorithms Non-orthogonal multiple access inter-cell interference co-ordination centralized radio access networks power map proportional fair scheduling communication system performance communication system planning optimization methods scheduling heterogeneous successive interference cancellation capability

来源：评论

学校读者我要写书评

暂无评论

User-Centric Resource scheduling for Dual-Connectivity Communications

引用

IEEE COMMUNICATIONS LETTERS 2021年第11期25卷 3659-3663页

作者： Cui, Haixia You, Fan South China Normal Univ Sch Elect & Informat Engn Guangzhou 510006 Peoples R China

In this letter, we study adaptive user-centric resource scheduling management for dual-connectivity (DC) heterogeneous networks with different numbers of user terminals (UTs) and different quality-of-service (QoS) requirements. We present a general framework for adaptively selecting the connection mode between UTs and base stations (BSs) to maximize energy efficiency. Based on this framework, we propose a user-centric resource scheduling algorithm to further increase network capacity, including traffic offloading among heterogeneous access technologies and available resource allocation, which results in less traffic congestion and better QoS of UTs. Simulation results demonstrate that our proposed algorithm can significantly improve energy efficiency and throughput over state-of-the-art schemes.

关键词： Bandwidth Base stations Resource management Uplink scheduling algorithms Optimization Heterogeneous networks Energy efficiency traffic offloading resource allocation user-centric dual connectivity (DC)

来源：评论

学校读者我要写书评

暂无评论

High-Definition High-Bandwidth DC/AC Power Conversion Through Binary Asymmetric Cascaded H-Bridges

引用

IEEE TRANSACTIONS ON POWER ELECTRONICS 2022年第11期37卷 13060-13069页

作者： Li, Zhongxi Goetz, Stefan M. Duke Univ Dept Elect & Comp Engn Durham NC 27708 USA

It is known that cascaded H-bridges with heterogenous module voltages can, in theory, generate exponentially-refined output voltage levels and achieve high output resolutions from only one power source. However, several critical considerations have yet to be thoroughly studied, namely the controllability of module voltages under any load conditions, and how to schedule the switching states optimally. We present a scheduling algorithm and topological modifications for binary asymmetric cascaded H-bridges (ACHBs) to achieve complete sensorless operation under any load conditions while using only one power source. The scheduling algorithm cancels the energy intake of floating modules per time frame, while low-power auxiliary dc/dc circuits, which naturally emerge from the bridge structure, can correct any voltage drift without the need for voltage sensors. Rigorous analyses reveal an inherent performance tradeoff between reference-tracing accuracy and floating module voltage controllability for ACHBs in general, for which the proposed algorithm always achieves the Pareto front. The proposed ACHB and the algorithm are tested on an experimental setup with six modules and 65 output levels (i.e., effectively six bits).

关键词： Switches Voltage control scheduling algorithms Inverters Controllability Redundancy Technological innovation Binary asymmetrical cascaded H-bridges (ACHBs) DC AC applications high definition

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：