A general scheduling framework (GSF) for independent tasks in computational Grid is proposed in this paper, which modeled by Petri net and located on the layer of Grid scheduler. Furthermore, a new mapping algorithm a...A general scheduling framework (GSF) for independent tasks in computational Grid is proposed in this paper, which modeled by Petri net and located on the layer of Grid scheduler. Furthermore, a new mapping algorithm aimed at time and cost is designed on the basis of this framework. The algorithm uses weighted average fuzzy applicability to express the matching degree between available machines and independent tasks. Some existent heuristic algorithms are tested in GSF, and the results of simulation and comparison not only show good flexibility and adaptability of GSF, but also prove that, given a certain aim, the new algorithm can consider the factors of time and cost as a whole and its performance is higher than those mentioned algorithms.展开更多
Researches related to wireless sensor networks primarily concentrate on Routing, Location Services, Data Aggregation and Energy Calculation Methods. Due to the heterogeneity of sensor networks using the web architectu...Researches related to wireless sensor networks primarily concentrate on Routing, Location Services, Data Aggregation and Energy Calculation Methods. Due to the heterogeneity of sensor networks using the web architecture, cross layer mechanism can be implemented for integrating multiple resources. Framework for Sensor Web using the cross layer scheduling mechanisms in the grid environment is proposed in this paper. The resource discovery and the energy efficient data aggregation schemes are used to improvise the effective utilization capability in the Sensor Web. To collaborate with multiple resources environment, the grid computing concept is integrated with sensor web. Resource discovery and the scheduling schemes in the grid architecture are organized using the medium access control protocol. The various cross layer metrics proposed are Memory Awareness, Task Awareness and Energy Awareness. Based on these metrics, the parameters-Node Waiting Status, Used CPU Status, Average System Utilization, Average Utilization per Cluster, Cluster Usage per Hour and Node Energy Status are determined for the integrated heterogeneous WSN with sensor web in Grid Environment. From the comparative analysis, it is shown that sensor grid architecture with middleware framework has better resource awareness than the normal sensor network architectures.展开更多
With the rapid development and popularization of 5G and the Internetof Things, a number of new applications have emerged, such as driverless cars.Most of these applications are time-delay sensitive, and some deficienc...With the rapid development and popularization of 5G and the Internetof Things, a number of new applications have emerged, such as driverless cars.Most of these applications are time-delay sensitive, and some deficiencies werefound during data processing through the cloud centric architecture. The data generated by terminals at the edge of the network is an urgent problem to be solved atpresent. In 5 g environments, edge computing can better meet the needs of lowdelay and wide connection applications, and support the fast request of terminalusers. However, edge computing only has the edge layer computing advantage,and it is difficult to achieve global resource scheduling and configuration, whichmay lead to the problems of low resource utilization rate, long task processingdelay and unbalanced system load, so as to lead to affect the service quality ofusers. To solve this problem, this paper studies task scheduling and resource collaboration based on a Cloud-Edge-Terminal collaborative architecture, proposes agenetic simulated annealing fusion algorithm, called GSA-EDGE, to achieve taskscheduling and resource allocation, and designs a series of experiments to verifythe effectiveness of the GSA-EDGE algorithm. The experimental results showthat the proposed method can reduce the time delay of task processing comparedwith the local task processing method and the task average allocation method.展开更多
This paper focuses on solving a problem of improving system robustness and the efficiency of a distributed system at the same time. Fault tolerance with active replication and load balancing techniques are used. The p...This paper focuses on solving a problem of improving system robustness and the efficiency of a distributed system at the same time. Fault tolerance with active replication and load balancing techniques are used. The pros and cons of both techniques are analyzed, and a novel load balancing framework for fault tolerant systems with active replication is presented. Hierarchical architecture is described in detail. The framework can dynamically adjust fault tolerant groups and their memberships with respect to system loads. Three potential task scheduler group selection methods are proposed and simulation tests are made. Further analysis of test data is done and helpful observations for system design are also pointed out, including effects of task arrival intensity and task set size, relationship between total task execution time and single task execution time.展开更多
电制氢(power to hydrogen,P2H)技术是“双碳”目标背景下实现电力脱碳的关键技术。受碱性电解池(alkaline electrolysis cell,AEC)单机容量的限制,P2H厂(站)利用AEC的可拓展性形成大容量多机集群系统。然而,现有调度运行框架研究难以保...电制氢(power to hydrogen,P2H)技术是“双碳”目标背景下实现电力脱碳的关键技术。受碱性电解池(alkaline electrolysis cell,AEC)单机容量的限制,P2H厂(站)利用AEC的可拓展性形成大容量多机集群系统。然而,现有调度运行框架研究难以保证P2H多机集群系统的性能最优。为此,提出了一种考虑效率优化的P2H多机集群系统调度运行框架,旨在提升系统效率并优化AEC间的功率分配。首先,分析了温度与负载率对产氢效率的影响,建立了包含辅机系统的P2H效率模型。然后,在此基础上构建了P2H多机集群系统效率优化模型,将其线性化为混合整数规划问题求解最优效率下各机组运行功率,并进一步提出P2H多机集群系统的优化调度模型。最后,基于IEEE 33节点配电网系统进行算例分析。结果表明,所提调度运行框架能降低系统运行成本,提升系统产氢效率和产氢量。研究成果可为P2H多机集群系统高效经济运行提供参考。展开更多
基金Project (60433020) supported by the National Natural Science Foundation of China project supported by the Postdoctor-al Science Foundation of Central South University
文摘A general scheduling framework (GSF) for independent tasks in computational Grid is proposed in this paper, which modeled by Petri net and located on the layer of Grid scheduler. Furthermore, a new mapping algorithm aimed at time and cost is designed on the basis of this framework. The algorithm uses weighted average fuzzy applicability to express the matching degree between available machines and independent tasks. Some existent heuristic algorithms are tested in GSF, and the results of simulation and comparison not only show good flexibility and adaptability of GSF, but also prove that, given a certain aim, the new algorithm can consider the factors of time and cost as a whole and its performance is higher than those mentioned algorithms.
文摘Researches related to wireless sensor networks primarily concentrate on Routing, Location Services, Data Aggregation and Energy Calculation Methods. Due to the heterogeneity of sensor networks using the web architecture, cross layer mechanism can be implemented for integrating multiple resources. Framework for Sensor Web using the cross layer scheduling mechanisms in the grid environment is proposed in this paper. The resource discovery and the energy efficient data aggregation schemes are used to improvise the effective utilization capability in the Sensor Web. To collaborate with multiple resources environment, the grid computing concept is integrated with sensor web. Resource discovery and the scheduling schemes in the grid architecture are organized using the medium access control protocol. The various cross layer metrics proposed are Memory Awareness, Task Awareness and Energy Awareness. Based on these metrics, the parameters-Node Waiting Status, Used CPU Status, Average System Utilization, Average Utilization per Cluster, Cluster Usage per Hour and Node Energy Status are determined for the integrated heterogeneous WSN with sensor web in Grid Environment. From the comparative analysis, it is shown that sensor grid architecture with middleware framework has better resource awareness than the normal sensor network architectures.
基金supported by the Social Science Foundation of Hebei Province(No.HB19JL007)the Education technology Foundation of the Ministry of Education(No.2017A01020)the Natural Science Foundation of Hebei Province(F2021207005).
文摘With the rapid development and popularization of 5G and the Internetof Things, a number of new applications have emerged, such as driverless cars.Most of these applications are time-delay sensitive, and some deficiencies werefound during data processing through the cloud centric architecture. The data generated by terminals at the edge of the network is an urgent problem to be solved atpresent. In 5 g environments, edge computing can better meet the needs of lowdelay and wide connection applications, and support the fast request of terminalusers. However, edge computing only has the edge layer computing advantage,and it is difficult to achieve global resource scheduling and configuration, whichmay lead to the problems of low resource utilization rate, long task processingdelay and unbalanced system load, so as to lead to affect the service quality ofusers. To solve this problem, this paper studies task scheduling and resource collaboration based on a Cloud-Edge-Terminal collaborative architecture, proposes agenetic simulated annealing fusion algorithm, called GSA-EDGE, to achieve taskscheduling and resource allocation, and designs a series of experiments to verifythe effectiveness of the GSA-EDGE algorithm. The experimental results showthat the proposed method can reduce the time delay of task processing comparedwith the local task processing method and the task average allocation method.
文摘This paper focuses on solving a problem of improving system robustness and the efficiency of a distributed system at the same time. Fault tolerance with active replication and load balancing techniques are used. The pros and cons of both techniques are analyzed, and a novel load balancing framework for fault tolerant systems with active replication is presented. Hierarchical architecture is described in detail. The framework can dynamically adjust fault tolerant groups and their memberships with respect to system loads. Three potential task scheduler group selection methods are proposed and simulation tests are made. Further analysis of test data is done and helpful observations for system design are also pointed out, including effects of task arrival intensity and task set size, relationship between total task execution time and single task execution time.
