• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.17 no.11
• /
• pp.1206-1218
• /
• 1992

Token ring systems which control to switch the data stream of networks by passing the token have been widely used to medium access controls in many ring or bus topology LANs. The system could be modeled for analysis as single-server-multi-queue system of the cyclic service method. These concepts could be expanded to multi-token ring systems interconnected with single ring consisting of bridges implemented simply to be stored and transmitted. In the proposal for the performance analysis of the interconnected token ring system, in has been assumed M/G/1 queueing model that frame arrivals are the Poisson process at each station queue and frame sizes are independently and identically distributed. And the average time delays were analyzed mathematically for arbitrary frame transferred from source station to destination area. The time delay of the frame transmission could be explained as the sum of the average time which the token passed from arbitrary position to source station, such as the waiting time in the source station transferring the previous arrival frames, and the propagation time from source station to interdestinated point. These delays were given as the sum of the duration from inner and outer bridge queues, the time delays from inner and outer bridge queues, and the time from outer bridge queue to destination station. These results were investigated by varing parameters effected to total time delays. In the results, those factors to be effected to dominant the total time delays were increased were in the cases of the high arrival rates and the high ration of destination of the other outerring. The system were shown the time delays increased exponentially in spite of the priority service policy. In order to decreasing the number of outerrings and increasing the number of nodes in backbone relatively, so the systems could be decreased the total delay in the interconnected token ring system.

• Proceedings of the KIEE Conference
• /
• 1991.07a
• /
• pp.748-752
• /
• 1991

Tree network consisting of communicating processors is considered. The objective is to minimize the computation time by distributing the processing load to other nodes. The effect of the order of load distribution on the processing time is addressed. An algorithm which optimally determines the order of load distribution is developed. It is shown that the order depends only on the channel capacity between nodes but not on the computing capability of each node.

• International Journal of Control, Automation, and Systems
• /
• v.2 no.2
• /
• pp.182-188
• /
• 2004

This paper presents a design method of delay-dependent control for T-S fuzzy systems with time delays. Based on parallel distributed compensation (PDC) and a descriptor model transformation of the system, a delay-dependent control is utilized. An appropriate Lyapunov-Krasovskii functional is chosen for delay-dependent stability analysis. A sufficient condition for delay-dependent control is represented in terms of linear matrix inequalities (LMIs).

• Journal of Platform Technology
• /
• v.11 no.4
• /
• pp.71-79
• /
• 2023

In this paper, we analyze the characteristics of machine learning workloads and, based on them, propose a distributed in-memory caching technique to improve the performance of machine learning workloads. The core of machine learning workload is model training, and model training is a computationally intensive task. Performing machine learning workloads in a Kubernetes-based cloud environment in which the computing framework and storage are separated can effectively allocate resources, but delays can occur because IO must be performed through network communication. In this paper, we propose a distributed in-memory caching technique to improve the performance of machine learning workloads performed in such an environment. In particular, we propose a new method of precaching data required for machine learning workloads into the distributed in-memory cache by considering Kubflow pipelines, a Kubernetes-based machine learning pipeline management tool.

• Proceedings of the KIEE Conference
• /
• 2005.10b
• /
• pp.246-248
• /
• 2005

The Controller Area Network (CAN) is being widely used for real-time control application and small-scale distributed computer controller systems. However, CAN may exhibit unfair behavior under heavy traffic conditions. When there are both high and low priority messages ready for transmission, the proposed precedence priority filtering method allows one low priority message to be exchanged between any two adjacent higher priority messages. In this way, the length of each transmission delays is upper bounded. These procedures are implemented as local controllers for the ISHURO (Inha Semyung Humanoid RObot).

• International Journal of Precision Engineering and Manufacturing
• /
• v.3 no.3
• /
• pp.12-20
• /
• 2002

As numerous sensors and actuators are used in many automated systems, various industrial networks are adopted for real-time distributed control. In order to take advantages of the networking, however, the network implementation should be carefully designed to satisfy real-time requirements considering network induced delays. This paper presents an implementation scheme of a networked control system via Profibus-DP network fur real-time distributed control. More specifically, the effect of the network induced delay on the control performance is evaluated on a Profibus-DP testbed. Also, two conventional PID gain tuning methods are slightly modified fur fouling controllers fur the networked control system. With appropriate choices for gains, it is shown that the networked control system can perform almost as well as the traditional control system.

• ETRI Journal
• /
• v.34 no.5
• /
• pp.771-774
• /
• 2012

Typical channel allocation algorithms for secondary users do not include processes to reduce the frequency of switching from one channel to another caused by random interruptions by primary users, which results in high packet drops and delays. In this letter, with the purpose of decreasing the number of switches made between channels, we propose a nonparametric channel allocation algorithm that uses robust kernel density estimation to effectively schedule idle channel resources. Experiment and simulation results demonstrate that the proposed algorithm outperforms both random and parametric channel allocation algorithms in terms of throughput and packet drops.

• Journal of Advanced Marine Engineering and Technology
• /
• v.28 no.4
• /
• pp.603-610
• /
• 2004

Time delays are included in most of actual systems, and some of which are shown as time varying. To analyze the time varying time delay system in the time domain. a useful delay module to express the function as a tool is much helpful to get corresponding outputs under given conditions. A method is proposed here to design the algorithm of time delay module for simulation or control purposes, including the problems of initializing and reallocating data in buffer. After classifying the time varying time delay into the distributed mode and lumped mode, an object to describe delay module is configured and tested under the defined input signal and given time delay variation. The simulation results show that the output of module matches reasonably with the case of real system.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• v.18 no.4
• /
• pp.317-335
• /
• 2014

In this work, we investigate the global stability analysis of a viral infection model with antibody immune response. The incidence rate is given by a general function of the populations of the uninfected target cells, infected cells and free viruses. The model has been incorporated with two types of intracellular distributed time delays to describe the time required for viral contacting an uninfected cell and releasing new infectious viruses. We have established a set of conditions on the general incidence rate function and determined two threshold parameters $R_0$ (the basic infection reproduction number) and $R_1$ (the antibody immune response activation number) which are sufficient to determine the global dynamics of the model. The global asymptotic stability of the equilibria of the model has been proven by using Lyapunov theory and applying LaSalle's invariance principle.

• Journal of the Korean Society for Precision Engineering
• /
• v.18 no.5
• /
• pp.120-129
• /
• 2001

As many sensors and actuators are used in many automated systems, various industrial networks are adopted for real-time distributed control. In order to take advantages of the networking, however, the network implementation should be carefully designed to satisfy real-time requirements considering network delays. This paper presents the implementation of a remote feedback control system via Profibus-DP net work for real-time distributed control More specifically, the effect of the network delay on the control performance evaluated on Profibus-DP testbed. Also, the traditional PID gain tuning methods are used to demonstrate the feasibility of the remote feedback control.