Hierarchical Scheduling Scheme Research Articles

A high-performance convolution block oriented accelerator for MBConv-Based CNNs

Convolution Neural Network (CNN) models have shown their dominance in computer vision tasks. Recently, a special convolution block, named MBConv block or inverted residual block, is proposed to construct CNNs to meet the real-time requirements on resource-constrained edge-computing platforms. The MBConv block is first proposed by MobileNetV2 and has been widely used to construct lightweight CNNs. However, the MBConv block brings new challenges to the structure of the computing engine, the bandwidth requirement of off-chip memory and the demand for on-chip memory when designing hardware accelerators. In this paper, a convolution Block Oriented Accelerator (BOA) is proposed for the inference of CNNs constructed on MBConv blocks. In BOA, the MBConv-based CNNs are performed block by block using a Block-Based Engine which consists of dedicated computing units for each layer of the MBConv block. To reduce both the bandwidth requirement of off-chip memory and the demand for on-chip memory, a two-level data flow optimization and an amortized weight loading method are proposed. Furthermore, a hierarchical scheduling scheme is proposed to improve the performance and flexibility so that BOA can guarantee all units running in parallel and support various MBConv-based CNNs. Finally, we deploy BOA on Xilinx VC709. We evaluate the accelerator on ImageNet for image classification. The results show that BOA can perform various MBConv-based CNNs and achieve 1.28x - 7.75x speedup on inference latency.

Multi-timescale and control-perceptive scheduling approach for flexible operation of power plant-carbon capture system

A carbon-capture system is required for coal-fired power plants to meet environmental regulation and provide support for grid penetration of renewable energy sources, when the integrated power plant-carbon capture system must operate in a highly flexible manner to accommodate the need for intermittent renewable power. To this end, this paper proposes a hierarchical scheduling scheme for the operation of integrated power plant-carbon capture system to fully exploit the decarbonization and flexibility of carbon capture technology in multiple timescales. The upper layer of the hierarchy implements a day-ahead stochastic scheduling to provide hourly operating instructions in the next 24 h so that the overall economic performance of the plant can be best achieved, including fuel consumption, load following and carbon trading. A novel index is then introduced to reflect the power adjustment contribution of carbon capture in the intraday scheduling layer and enhance the load ramping performance of power plant by flexible operation of the post-combustion carbon capture in minute timescale. In the lower layer of the hierarchy, control-perceptive scheduling based on close-loop dynamic model is proposed to coordinate the scheduling and underlying control, making the scheduling feasible for the operating practice. In addition, a new control system, which uses the reboiler steam flowrate to control the CO2 production rate and solvent circulation flowrate to control reboiler temperature is proposed to achieve superior flexibility support role of carbon capture. The case study shows that the proposed scheduling approach can improve the load tracking performance by 36.7% with satisfactory performance in carbon capture, which verifies the efficacy of the proposed scheduling approach in the power system. The impact of the proposed control system is also investigated, providing a broad insight on the flexible operation of carbon capture technology integrated with the power plant.

Low Complexity Hierarchical Scheduling for Diverse Datacenter Jobs

Datacenter jobs in a private cloud are often generated by different departments or groups of a business or other organization. In order to reflect the organizational structures of datacenter workloads, it is desirable for the datacenter scheduler to support hierarchical scheduling. The state-of-the-art hierarchical scheduling scheme, hierarchical dominant resource fairness (H-DRF), however, is expensive to implement. Specifically, the worst-case time complexity of H-DRF is $O(q^{2})$ , where $q$ is dominated by the number of jobs in the hierarchy, which is a large complexity in a practical large-scale datacenter. In this letter, we design a new online hierarchical fair scheduling scheme, multi-resource collapsed hierarchies (MCH), in the way of converting the hierarchy into a weighted flat tree. MCH takes at most $O(q)$ time and is simple enough to implement in practice. Simulations driven by Google cluster traces show that MCH consumes a nearly sqrt of the total run time of H-DRF, improving significantly in saving the computation time.

Hierarchical Scheduling Scheme for AC/DC Hybrid Active Distribution Network Based on Multi-Stakeholders

This paper presents a hierarchical multi-stage scheduling scheme for the AC/DC hybrid active distribution network (ADN). The load regulation center (LRC) is considered in the developed scheduling strategy, as well as the AC and DC sub-network operators. They are taken to be different stakeholders. To coordinate the interests of all stakeholders, a two-level optimization model is established. The flexible loads are dispatched by LRC in the upper-level optimization model, the objective of which is minimizing the loss of the entire distribution network. The lower-level optimization is divided into two sub-optimal models, and they are carried out to minimize the operating costs of the AC/DC sub-network operators respectively. This two-level model avoids the difficulty of solving multi-objective optimization and can clarify the role of various stakeholders in the system scheduling. To solve the model effectively, a discrete wind-driven optimization (DWDO) algorithm is proposed. Then, considering the combination of the proposed DWDO algorithm and the YALMIP toolbox, a hierarchical optimization algorithm (HOA) is developed. The HOA can obtain the overall optimization result of the system through the iterative optimization of the upper and lower levels. Finally, the simulation results verify the effectiveness of the proposed scheduling scheme.

A Hierarchical Scheduling Scheme in WSNs Based on Node-Failure Pretreatment

One of the most important challenges in designing wireless sensor networks is how to recover a broken network within a very short time as the active nodes are failed or out-of-energy. Focusing on this problem, a hierarchical scheduling scheme based on node-failure pretreatment is proposed, in which the global optimization method is used to find the minimum connected tree and the local multilayer recovery algorithm is used to find a candidate sensor node instead of the failed one. Three highlights of this scheme are as follows: (1) The importance of sensor nodes is defined in terms of their locations in minimum connected tree and coverage acreage. (2) The neighborhood radius of failed sensor node varies with its importance, and then its candidate-node set is dynamically constructed. (3) A novel multilayer recovery strategy including node recovery and regional recovery is presented. Simulation results show that hierarchical scheduling scheme finds the optimal candidate sensor node in less time to make the repaired network with lowest energy consumption. Though the less sensor nodes are activated, the network lifetime is slightly shorter. Moreover, this scheme can be applied in the problem that the communication radius of sensor node is less than two times of its sensing radius.

An adaptive and hierarchical task scheduling scheme for multi-core clusters

• An adaptive and hierarchical task scheduling scheme (AHS) is proposed. • Work-sharing is used in conjunction with work-stealing. • An initial partitioning is performed with respect to the pattern of task parallelism . • A practical implementation of AHS is described. • The theoretical, simulation and experimental studies of AHS are presented. Work-stealing and work-sharing are two basic paradigms for dynamic task scheduling. This paper introduces an adaptive and hierarchical task scheduling scheme (AHS) for multi-core clusters, in which work-stealing and work-sharing are adaptively used to achieve load balancing. Work-stealing has been widely used in task-based parallel programing languages and models, especially on shared memory systems. However, high inter-node communication costs hinder work-stealing from being directly performed on distributed memory systems. AHS addresses this issue with the following techniques: (1) initial partitioning, which reduces the inter-node task migrations; (2) hierarchical scheduling scheme, which performs work-stealing inside a node before going across the node boundary and adopts work-sharing to overlap computation and communication at the inter-node level; and (3) hierarchical and centralized control for inter-node task migration, which improves the efficiency of victim selection and termination detection. We evaluated AHS and existing work-stealing schemes on a 16-nodes multi-core cluster. Experimental results show that AHS outperforms existing schemes by 11–21.4%, for the benchmarks studied in this paper.

QoS-aware adaptation for satellite multimedia broadcasting via hierarchical packet scheduling

With the unique broadcast nature and ubiquitous coverage of satellite networks, the synergy between satellite and terrestrial networks provides opportunities for delivering wideband services to a wide range of audiences over extensive geographical areas. This article concerns the optimization techniques pertinent to packet scheduling to facilitate multimedia content delivery over the satellite with a return channel via a terrestrial network. We propose a novel hierarchical packet scheduling scheme that allocates the resources at different parts of the network in response to network dynamics and link variations while under the system power/resource constraints. Simulations prove that the HPS scheme can effectively improve the end-to-end performance and resource utilization with desirable scalability and fairness features.

Two-level hierarchical scheduling scheme of hybrid tasks for software CNC system

For the scheduling of hybrid tasks for software CNC,a two-level hierarchical scheduling scheme is presented.The time-triggered mode and event-triggered mode are adopted for the real-time periodic tasks and real-time aperiodic tasks respectively. Firstly,the task model is established and the related concepts are defined.Then,the structure of scheduling scheme is proposed,and scheduling algorithm,schedulability and tick-interval are deeply researched.Subsequently,the scheduling scheme is adopted for the software CNC based on Windows CE.NET,and software CNC structure and task division are analyzed.Finally,the experimental verification of two-level hierarchical scheduling scheme is carried out and the results demonstrate that the proposed method can ensure the determination of real-time tasks,and improve the performance of software CNC system.

An OFDM-TDMA/SA MAC Protocol with QoS Constraints for Broadband Wireless LANs

Orthogonal frequency division multiplexing (OFDM) is an important technique to support high speed transmission of broadband traffic in wireless networks, especially broadband wireless local area networks (LANs). Based on OFDM, a new multiple access scheme, called OFDM-TDMA with subcarrier allocation (OFDM-TDMA/SA), is proposed in this paper. It provides more flexibility in resource allocation than other multiple access schemes such as OFDM-TDMA, OFDM-frequency division multiple access (OFDM-FDMA), and orthogonal frequency division multiple access (OFDMA). With OFDM-TDMA/SA, a medium access control (MAC) is designed for broad-band wireless LANs. It optimizes bit allocation in subcarriers so that maximum bits are transmitted in each OFDM symbol under a frequency selective fading environment. The OFDM-TDMA/SA MAC protocol also supports three classes of traffic such as guaranteed, controlled-load, and best effort services. Based on the optimum subcarrier bit-allocation algorithm and considering heterogeneous QoS constraints of multimedia traffic, a hierarchical scheduling scheme is proposed to determine the subcarriers and time slots in which a mobile terminal can transmit packets. In such a way, the OFDM-TDMA/SA MAC protocol significantly increases system throughput in a frequency selective fading environment and guarantees QoS of multimedia traffic. Computer simulation is carried out to evaluate the performance of the OFDM-TDMA/SA MAC protocol. Results show that the new MAC protocol outperforms other MAC protocols for OFDM-based wireless LANs.

A unified partitioning and scheduling scheme for mapping multi-stage regular iterative algorithms onto processor arrays

This paper addresses the partitioning and scheduling problems in mapping multi-stage regular iterative algorithms onto fixed size distributed memory processor arrays. We first propose a versatile partitioning model which provides a unified framework to integrate various partitioning schemes such as "locally sequential globally parallel", "locally parallel globally sequential" and "multi-projection". To alleviate the run time data migration overhead--a crucial problem to the mapping of multi-stage algorithms, we further relax the widely adopted atomic partitioning constraint in our model such that a more flexible partitioning scheme can be achieved. Based on this unified partitioning model, a novel hierarchical scheduling scheme which applies separate schedules at different processor hierarchies is then developed. The scheduling problem is then formulated into a set of ILP problem and solved by the existing software package for optimal solutions. Examples indicate that our partitioning model is a superset of the existing schemes and the proposed hierarchical scheduling scheme can outperform the conventional one-level linear schedule.

Hydro Coordination in an Energy Management System

Coordination of hydro resources over the operational planning period of one year is a challenging mathematical optimization problem. The application of this operational planning strategy through weekly, hourly, and real-time scheduling of hydro units increases the complexity of the hydro coordination problem further.For a computer based water management system (WMS), a hierarchical scheduling scheme is proposed so that the yearly strategy can be meaningfully translated into weekly and real-time decisions for loading the hydro units. In this process WMS is decomposed into the major functional units of yearly hydro schedule, weekly hydro schedule, real-time scheduling and control, and water use monitor. The paper addresses the problems associated with each function, and the coordination of these functional units in an EMS.

Performance of Hierarchical Production Scheduling Policy

The performance of Kimemia and Gershwin's hierarchical scheduling scheme for flexible manufacturing systems, as enhanced by Gershwin, Akella, and Choong, is described. This method calculates times at which to dispatch parts into a system in a way that limits the disruptive effects of such disturbances as machine failures. Simulation results based on a detailed model of an IBM printed circuit card assembly facility are presented. Comparisons are made with other policies and the hierarchical policy is shown to be superior.

Fuel Resource Scheduling, Part I-Overview of an Energy Management Problem

This paper at an overview level discusses the problems of fuel and resource scheduling (FRS) in an energy management system (EMS). For a computer-based FRS, a hierarchical scheduling scheme is proposed so that long-term strategy over a year can be meaningfully translated into realtime decisions for loading the units. In this process FRS has been decomposed into individual problems of resource budgeting for long-term fuel scheduling, resource scheduling for short-term fuel scheduling, and real-time fuel scheduling functions of real-time resource scheduling and resource constrained economic dispatch. The paper addresses the data requirements and the technical problems associated with each individual function; and the coordination of these four functions in an EMS environment. Fuel resource scheduling (FRS) optimizes long-term fuel consumption based upon load and generating unit characteristics, and resource availability constraints. FRS translates the optimum long-term strategy to real-time (hours to minutes) constraints to be used by the economic dispatch for the realtime allocation of system generation requirement among the units. FRS covers a wide range of time periods (from one year to minutes) and involves a large number of constraints and variables. This makes the fuel constrained generation dispatch and control process extremely difficult for system dispatchers even with the aid of computers. This paper addresses the problems associated with FRS as an integral part of an energy management system.

Fuel Resource Scheduling, Part I: Overview of an Energy Management Problem

This paper at an overview level discusses the problems of fuel and resource scheduling (FRS) in an energy management system (EMS). For a computer-based FRS, a hierarchical scheduling scheme is proposed so that long-term strategy over a year can be meaningfully translated into realtime decisions for loading the units. In this process FRS has been decomposed into individual problems of resource budgeting for long-term fuel scheduling, resource scheduling for short-term fuel scheduling, and real-time fuel scheduling functions of real-time resource scheduling and resource constrained economic dispatch. The paper addresses the data requirements and the technical problems associated with each individual function; and the coordination of these four functions in an EMS environment. Fuel resource scheduling (FRS) optimizes long-term fuel consumption based upon load and generating unit characteristics, and resource availability constraints. FRS translates the optimum long-term strategy to real-time (hours to minutes) constraints to be used by the economic dispatch for the realtime allocation of system generation requirement among the units. FRS covers a wide range of time periods (from one year to minutes) and involves a large number of constraints and variables. This makes the fuel constrained generation dispatch and control process extremely difficult for system dispatchers even with the aid of computers. This paper addresses the problems associated with FRS as an integral part of an energy management system.