The high-performance computing paradigm needs high-speed switching fabrics to meet the heavy traffic generated by their applications.These switching fabrics are efficiently driven by the deployed scheduling algorithms...The high-performance computing paradigm needs high-speed switching fabrics to meet the heavy traffic generated by their applications.These switching fabrics are efficiently driven by the deployed scheduling algorithms.In this paper,we proposed two scheduling algorithms for input queued switches whose operations are based on ranking procedures.At first,we proposed a Simple 2-Bit(S2B)scheme which uses binary ranking procedure and queue size for scheduling the packets.Here,the Virtual Output Queue(VOQ)set with maximum number of empty queues receives higher rank than other VOQ’s.Through simulation,we showed S2B has better throughput performance than Highest Ranking First(HRF)arbitration under uniform,and non-uniform traffic patterns.To further improve the throughput-delay performance,an Enhanced 2-Bit(E2B)approach is proposed.This approach adopts an integer representation for rank,which is the number of empty queues in a VOQ set.The simulation result shows E2B outperforms S2B and HRF scheduling algorithms with maximum throughput-delay performance.Furthermore,the algorithms are simulated under hotspot traffic and E2B proves to be more efficient.展开更多
The sharing of operation and management information is required by smart grid.The amount of information transmitted in electric power communication systems will increase greatly in the near future.The integrated infor...The sharing of operation and management information is required by smart grid.The amount of information transmitted in electric power communication systems will increase greatly in the near future.The integrated information transmission technology is an important method for transmitting various kinds of information via the existing broadband channels and networks in power systems.To implement integrated information transmission in power systems,the problem about how to guarantee the quality of service(QoS)of the communication services must be dealt with.展开更多
It is well known, in queueing theory, that the system performance is greatly influenced by scheduling policy. No universal optimum scheduling strategy exists in systems where individual customer service demands are no...It is well known, in queueing theory, that the system performance is greatly influenced by scheduling policy. No universal optimum scheduling strategy exists in systems where individual customer service demands are not known a priori. However, if the distribution of job times is known, then the residual time (expected time remaining for a job), based on the service it has already received, can be calculated. Our particular research contribution is in exploring the use of this function to enhance system performance by increasing the probability that a job will meet its deadline. In a detailed discrete event simulation, we have tested many different distributions with a wide range of C2 and shapes, as well as for single and dual processor system. Results of four distributions are reported here. We compare with RR and FCFS, and find that in all distributions studied our algorithm performs best. In the study of the use of two slow servers versus one fast server, we have discovered that they provide comparable performance, and in a few cases the double server system does better.展开更多
Currently, the article analyzes the CAN bus＇s rule of priority＇s arbitration bit by bit without destroy. It elicits the conclusion that if static priority based on the affirmatory system model is used, the lower pri...Currently, the article analyzes the CAN bus＇s rule of priority＇s arbitration bit by bit without destroy. It elicits the conclusion that if static priority based on the affirmatory system model is used, the lower priority＇s messages will be delayed considerably more, even some data will be lost when the bus＇s bandwidth is widely used. The scheduling cannot be modified neither during the system when static priority is used. The dynamic priority promoting method and the math model of SQSA and SQMA are presented; it analyzes the model＇s rate of taking in and sending out in large quantities, the largest delay, the problems and solutions when using SQMA. In the end, it is confirmed that the method of improving dynamic priority has good performances on the network rate of taking in and sending out in large quantities, the average delay, and the rate of network usage by emulational experiments.展开更多
This paper introduces a multi-granularity locking model (MGL) for concurrency control in object-oriented database system briefiy, and presents a MGL model formally. Four lockingscheduling algorithms for MGL are propos...This paper introduces a multi-granularity locking model (MGL) for concurrency control in object-oriented database system briefiy, and presents a MGL model formally. Four lockingscheduling algorithms for MGL are proposed in the paper. The ideas of single queue scheduling(SQS) and dual queue scheduling (DQS) are proposed and the algorithm and the performance evaluation for these two scheduling are presented in some paper. This paper describes a new idea of thescheduling for MGL, compatible requests first (CRF). Combining the new idea with SQS and DQS,we propose two new scheduling algorithms called CRFS and CRFD. After describing the simulationmodel, this paper illustrates the comparisons of the performance among these four algorithms. Asshown in the experiments, DQS has better performance than SQS, CRFD is better than DQS, CRFSperforms better than SQS, and CRFS is the best one of these four scheduling algorithms.展开更多
One of the fundamental problems in parallel and distributed systems is deciding how to allocate jobs to processors. The goals of job scheduling in a parallel environment are to minimize the parallel execution time of ...One of the fundamental problems in parallel and distributed systems is deciding how to allocate jobs to processors. The goals of job scheduling in a parallel environment are to minimize the parallel execution time of a job and try to balance the user’s desire with the system’s desire. The users always want their jobs be completed as quickly as possible, while the system wants to service as many jobs as possible. In this paper, a dynamic job scheduling algorithm was introduced. This algorithm tries to utilize the information of a practical system to allocate the jobs more evenly. The communication time between the processor and scheduler is overlapped with the computation time of the processor. So the communication overhead can be little. The principle of scheduling the job is based on the desirability of each processor. The scheduler would not allocate a new job to a processor that is already fully utilized. The execution efficiency of the system will be increased. This algorithm also can be reused in other complex algorithms.展开更多
The problem of scheduling real time and non real time traffic in an ATM switch multiplexor when bursts of either type of traffic occur is studied. The scheduling algorithms studied are: Queue Length Threshold (QLT) an...The problem of scheduling real time and non real time traffic in an ATM switch multiplexor when bursts of either type of traffic occur is studied. The scheduling algorithms studied are: Queue Length Threshold (QLT) and Minimum Laxity Threshold (MLT). Analytic results based on Markov Chains are used. In addition the results are compared with an optimal (but impractical) scheduling determined via dynamic programming. Dynamic programming is used in this paper to show that MLT gives a near optimal performance trade off between real time and non real time traffic for constant arrival rates. The trade off QLT allows is not close to optimal. For non real time bursts MLT still gives a close to optimal trade off. For real time bursts the trade off MLT allows between real time and non real time traffic is not as close to optimal, but even where the MLT trade off is not near optimal, the QLT trade off is much worse than the MLT trade off.展开更多
Due to the fading characteristics of wireless channels and the burstiness of data traffic,how to deal with congestion in Ad-hoc networks with effective algorithms is still open and challenging.In this paper,we focus o...Due to the fading characteristics of wireless channels and the burstiness of data traffic,how to deal with congestion in Ad-hoc networks with effective algorithms is still open and challenging.In this paper,we focus on enabling congestion control to minimize network transmission delays through flexible power control.To effectively solve the congestion problem,we propose a distributed cross-layer scheduling algorithm,which is empowered by graph-based multi-agent deep reinforcement learning.The transmit power is adaptively adjusted in real-time by our algorithm based only on local information(i.e.,channel state information and queue length)and local communication(i.e.,information exchanged with neighbors).Moreover,the training complexity of the algorithm is low due to the regional cooperation based on the graph attention network.In the evaluation,we show that our algorithm can reduce the transmission delay of data flow under severe signal interference and drastically changing channel states,and demonstrate the adaptability and stability in different topologies.The method is general and can be extended to various types of topologies.展开更多
This paper proposes a Reinforcement learning(RL)algorithm to find an optimal scheduling policy to minimize the delay for a given energy constraint in communication system where the environments such as traffic arrival...This paper proposes a Reinforcement learning(RL)algorithm to find an optimal scheduling policy to minimize the delay for a given energy constraint in communication system where the environments such as traffic arrival rates are not known in advance and can change over time.For this purpose,this problem is formulated as an infinite-horizon Constrained Markov Decision Process(CMDP).To handle the constrained optimization problem,we first adopt the Lagrangian relaxation technique to solve it.Then,we propose a variant of Q-learning,Q-greedyUCB that combinesε-greedy and Upper Confidence Bound(UCB)algorithms to solve this constrained MDP problem.We mathematically prove that the Q-greedyUCB algorithm converges to an optimal solution.Simulation results also show that Q-greedyUCB finds an optimal scheduling strategy,and is more efficient than Q-learning withε-greedy,R-learning and the Averagepayoff RL(ARL)algorithm in terms of the cumulative regret.We also show that our algorithm can learn and adapt to the changes of the environment,so as to obtain an optimal scheduling strategy under a given power constraint for the new environment.展开更多
针对车路协同系统(cooperative vehicle infrastructure system,CVIS)中排队时延和功率间的平衡问题,提出一种联合队列年龄和信道感知的概率调度(joint age of queue and channel aware probabilistic scheduling,JAQCA-PS)策略,在给定...针对车路协同系统(cooperative vehicle infrastructure system,CVIS)中排队时延和功率间的平衡问题,提出一种联合队列年龄和信道感知的概率调度(joint age of queue and channel aware probabilistic scheduling,JAQCA-PS)策略,在给定功率约束下实现平均排队时延最小化。提出的策略设计了队列年龄(age of queue,AoQ)信息作为调度参数之一,并联合包到达信息、块衰落信道的状态信息建立跨层调度模型,将最小化平均排队时延问题表述为约束马尔可夫决策过程(constrained Markov decision process,CMDP);通过将非线性问题转化为等效的线性规划(linear programming,LP)问题求解,得到了JAQCA-PS策略下最优传输参数和最优排队时延-功率平衡的数值结果;经蒙特卡洛模拟验证了结果的准确性。展开更多
针对星载路由设备的发展趋势,即交互速率提升、带宽增加以及设计轻量化,本文提出了一种基于改进轮询的星载高速交换算法的设计方案。该方案采用基于改进RR(Round Robin,轮询)调度算法和PBPW(Priority-based Bandwidth Privilege with We...针对星载路由设备的发展趋势,即交互速率提升、带宽增加以及设计轻量化,本文提出了一种基于改进轮询的星载高速交换算法的设计方案。该方案采用基于改进RR(Round Robin,轮询)调度算法和PBPW(Priority-based Bandwidth Privilege with Weighting,基于优先级并带权重的带宽特权)算法相结合的二级队列调度算法。在第一级调度中,引入优先级轮询调度,确保高优先级的数据帧能够优先转发,同时开放了缓存共享机制,一定程度上避免了拥塞和资源的浪费;在第二级调度中,为每条链路分配了阈值,从而避免了某些链路因“饥饿”而无法得到服务,同时防止了其他链路的拥塞问题。与传统的基于FIFO(First In First Out,先进先出)的简单队列调度算法相比,这种改进的轮询调度机制显著提高了星载路由器的转发速率,并降低了转发时延。此外,通过轮询方式实现了对高优先级数据帧的优先转发,进一步优化了路由器的性能。展开更多
文摘The high-performance computing paradigm needs high-speed switching fabrics to meet the heavy traffic generated by their applications.These switching fabrics are efficiently driven by the deployed scheduling algorithms.In this paper,we proposed two scheduling algorithms for input queued switches whose operations are based on ranking procedures.At first,we proposed a Simple 2-Bit(S2B)scheme which uses binary ranking procedure and queue size for scheduling the packets.Here,the Virtual Output Queue(VOQ)set with maximum number of empty queues receives higher rank than other VOQ’s.Through simulation,we showed S2B has better throughput performance than Highest Ranking First(HRF)arbitration under uniform,and non-uniform traffic patterns.To further improve the throughput-delay performance,an Enhanced 2-Bit(E2B)approach is proposed.This approach adopts an integer representation for rank,which is the number of empty queues in a VOQ set.The simulation result shows E2B outperforms S2B and HRF scheduling algorithms with maximum throughput-delay performance.Furthermore,the algorithms are simulated under hotspot traffic and E2B proves to be more efficient.
文摘The sharing of operation and management information is required by smart grid.The amount of information transmitted in electric power communication systems will increase greatly in the near future.The integrated information transmission technology is an important method for transmitting various kinds of information via the existing broadband channels and networks in power systems.To implement integrated information transmission in power systems,the problem about how to guarantee the quality of service(QoS)of the communication services must be dealt with.
文摘It is well known, in queueing theory, that the system performance is greatly influenced by scheduling policy. No universal optimum scheduling strategy exists in systems where individual customer service demands are not known a priori. However, if the distribution of job times is known, then the residual time (expected time remaining for a job), based on the service it has already received, can be calculated. Our particular research contribution is in exploring the use of this function to enhance system performance by increasing the probability that a job will meet its deadline. In a detailed discrete event simulation, we have tested many different distributions with a wide range of C2 and shapes, as well as for single and dual processor system. Results of four distributions are reported here. We compare with RR and FCFS, and find that in all distributions studied our algorithm performs best. In the study of the use of two slow servers versus one fast server, we have discovered that they provide comparable performance, and in a few cases the double server system does better.
基金supported by the National Natural Science Foundation of China (50421703)the National Key Laboratory of Electrical Engineering of Naval Engineering University
文摘Currently, the article analyzes the CAN bus＇s rule of priority＇s arbitration bit by bit without destroy. It elicits the conclusion that if static priority based on the affirmatory system model is used, the lower priority＇s messages will be delayed considerably more, even some data will be lost when the bus＇s bandwidth is widely used. The scheduling cannot be modified neither during the system when static priority is used. The dynamic priority promoting method and the math model of SQSA and SQMA are presented; it analyzes the model＇s rate of taking in and sending out in large quantities, the largest delay, the problems and solutions when using SQMA. In the end, it is confirmed that the method of improving dynamic priority has good performances on the network rate of taking in and sending out in large quantities, the average delay, and the rate of network usage by emulational experiments.
文摘This paper introduces a multi-granularity locking model (MGL) for concurrency control in object-oriented database system briefiy, and presents a MGL model formally. Four lockingscheduling algorithms for MGL are proposed in the paper. The ideas of single queue scheduling(SQS) and dual queue scheduling (DQS) are proposed and the algorithm and the performance evaluation for these two scheduling are presented in some paper. This paper describes a new idea of thescheduling for MGL, compatible requests first (CRF). Combining the new idea with SQS and DQS,we propose two new scheduling algorithms called CRFS and CRFD. After describing the simulationmodel, this paper illustrates the comparisons of the performance among these four algorithms. Asshown in the experiments, DQS has better performance than SQS, CRFD is better than DQS, CRFSperforms better than SQS, and CRFS is the best one of these four scheduling algorithms.
基金National Natural Science Foundation of China( No.60 173 0 3 1)
文摘One of the fundamental problems in parallel and distributed systems is deciding how to allocate jobs to processors. The goals of job scheduling in a parallel environment are to minimize the parallel execution time of a job and try to balance the user’s desire with the system’s desire. The users always want their jobs be completed as quickly as possible, while the system wants to service as many jobs as possible. In this paper, a dynamic job scheduling algorithm was introduced. This algorithm tries to utilize the information of a practical system to allocate the jobs more evenly. The communication time between the processor and scheduler is overlapped with the computation time of the processor. So the communication overhead can be little. The principle of scheduling the job is based on the desirability of each processor. The scheduler would not allocate a new job to a processor that is already fully utilized. The execution efficiency of the system will be increased. This algorithm also can be reused in other complex algorithms.
文摘The problem of scheduling real time and non real time traffic in an ATM switch multiplexor when bursts of either type of traffic occur is studied. The scheduling algorithms studied are: Queue Length Threshold (QLT) and Minimum Laxity Threshold (MLT). Analytic results based on Markov Chains are used. In addition the results are compared with an optimal (but impractical) scheduling determined via dynamic programming. Dynamic programming is used in this paper to show that MLT gives a near optimal performance trade off between real time and non real time traffic for constant arrival rates. The trade off QLT allows is not close to optimal. For non real time bursts MLT still gives a close to optimal trade off. For real time bursts the trade off MLT allows between real time and non real time traffic is not as close to optimal, but even where the MLT trade off is not near optimal, the QLT trade off is much worse than the MLT trade off.
基金supported by Institute of Information & communications Technology Planning & Evaluation (IITP) grant funded by the Korea government(MSIT) (No.RS-2022-00155885, Artificial Intelligence Convergence Innovation Human Resources Development (Hanyang University ERICA))supported by the National Natural Science Foundation of China under Grant No. 61971264the National Natural Science Foundation of China/Research Grants Council Collaborative Research Scheme under Grant No. 62261160390
文摘Due to the fading characteristics of wireless channels and the burstiness of data traffic,how to deal with congestion in Ad-hoc networks with effective algorithms is still open and challenging.In this paper,we focus on enabling congestion control to minimize network transmission delays through flexible power control.To effectively solve the congestion problem,we propose a distributed cross-layer scheduling algorithm,which is empowered by graph-based multi-agent deep reinforcement learning.The transmit power is adaptively adjusted in real-time by our algorithm based only on local information(i.e.,channel state information and queue length)and local communication(i.e.,information exchanged with neighbors).Moreover,the training complexity of the algorithm is low due to the regional cooperation based on the graph attention network.In the evaluation,we show that our algorithm can reduce the transmission delay of data flow under severe signal interference and drastically changing channel states,and demonstrate the adaptability and stability in different topologies.The method is general and can be extended to various types of topologies.
基金This work was supported by the research fund of Hanyang University(HY-2019-N)This work was supported by the National Key Research&Development Program 2018YFA0701601.
文摘This paper proposes a Reinforcement learning(RL)algorithm to find an optimal scheduling policy to minimize the delay for a given energy constraint in communication system where the environments such as traffic arrival rates are not known in advance and can change over time.For this purpose,this problem is formulated as an infinite-horizon Constrained Markov Decision Process(CMDP).To handle the constrained optimization problem,we first adopt the Lagrangian relaxation technique to solve it.Then,we propose a variant of Q-learning,Q-greedyUCB that combinesε-greedy and Upper Confidence Bound(UCB)algorithms to solve this constrained MDP problem.We mathematically prove that the Q-greedyUCB algorithm converges to an optimal solution.Simulation results also show that Q-greedyUCB finds an optimal scheduling strategy,and is more efficient than Q-learning withε-greedy,R-learning and the Averagepayoff RL(ARL)algorithm in terms of the cumulative regret.We also show that our algorithm can learn and adapt to the changes of the environment,so as to obtain an optimal scheduling strategy under a given power constraint for the new environment.
文摘针对车路协同系统(cooperative vehicle infrastructure system,CVIS)中排队时延和功率间的平衡问题,提出一种联合队列年龄和信道感知的概率调度(joint age of queue and channel aware probabilistic scheduling,JAQCA-PS)策略,在给定功率约束下实现平均排队时延最小化。提出的策略设计了队列年龄(age of queue,AoQ)信息作为调度参数之一,并联合包到达信息、块衰落信道的状态信息建立跨层调度模型,将最小化平均排队时延问题表述为约束马尔可夫决策过程(constrained Markov decision process,CMDP);通过将非线性问题转化为等效的线性规划(linear programming,LP)问题求解,得到了JAQCA-PS策略下最优传输参数和最优排队时延-功率平衡的数值结果;经蒙特卡洛模拟验证了结果的准确性。
文摘针对星载路由设备的发展趋势,即交互速率提升、带宽增加以及设计轻量化,本文提出了一种基于改进轮询的星载高速交换算法的设计方案。该方案采用基于改进RR(Round Robin,轮询)调度算法和PBPW(Priority-based Bandwidth Privilege with Weighting,基于优先级并带权重的带宽特权)算法相结合的二级队列调度算法。在第一级调度中,引入优先级轮询调度,确保高优先级的数据帧能够优先转发,同时开放了缓存共享机制,一定程度上避免了拥塞和资源的浪费;在第二级调度中,为每条链路分配了阈值,从而避免了某些链路因“饥饿”而无法得到服务,同时防止了其他链路的拥塞问题。与传统的基于FIFO(First In First Out,先进先出)的简单队列调度算法相比,这种改进的轮询调度机制显著提高了星载路由器的转发速率,并降低了转发时延。此外,通过轮询方式实现了对高优先级数据帧的优先转发,进一步优化了路由器的性能。
