• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2001.06a
• /
• pp.109-112
• /
• 2001

기존의 K-means 알고리즘은 학습벡터가 단일군집에 할당되는 방법이 crisp 이므로 다른 군집에 할당될 확률을 무시하게 된다. 따라서 군집화 작업과 관련하여 반복적인 코드북 설계 과정에서 각 학습벡터를 다중 군집으로 할당하는 Fuzzy c-means를 사용한다. 또한 Fuzzy c-means 알고리즘의 학습과정에서 구해지는 각 클래스 의 프로토타입에 가중치를 곱하여 다음 학습의 프로토타입으로 사용함으로써 Fuzzy c-means 알고리즘 적용 결과 얻어지는 코트북의 성능을 기존 알고리즘과 비교하여 개선된 Fuzzy c-means 알고리즘을 찾기 위한 근거를 마련한다.

• The Transactions of the Korea Information Processing Society
• /
• v.5 no.5
• /
• pp.1151-1161
• /
• 1998

In this paper, we are considering a heterogeneous processor system in which each processor may have different performance and reliability characteristics. In other to fully utilize this diversity of processing power it is advantageous to assign the program modules of a distributed program to the processors in such a way that the execution time of the entire program is minimized. This assignment of tasks to processors to maximize performance is commonly called load balancing, since the overloaded processors can perform their own processing with the performance degradation. For the task assignment problem, we propose a new objective function which formulates this imbalancing cost. Thus the task assignment problem is to be carried out so that each module is assigned to a processor whose capabilities are most appropriate for the module, and the total cost is minimized that sum of inter-processor communication cost and execution cost and imbalance cost of the assignment. To find optimal assignment is known to be NP-hard, and thus we proposed an efficient heuristic algorithm with time complexity $O(n^2m)$ in case of m task modules and n processors.

• Proceedings of the Korean Information Science Society Conference
• /
• 1998.10a
• /
• pp.136-138
• /
• 1998

작업들이 자원을 공유하는 경우 예측하기 어려운 지연시간이 발생한다. 다중 프로세서 시스템에서의 자원공유로 인한 지연시간은 더욱 예측하기 어렵다. 실기간 시스템의 스케줄 가능성 검사를 위해서는 이러한 지연시간을 정확히 예측해야한다. 선점가능한 우선순위 구동 CPU 스케줄링 알고리즘에 의해서 다른 우선순위의 작업과의 동기화는 우선순위 역전 문제를 야기한다. 본 논문에서는 다중 프로세서에서의 동기화 프로토콜을 제안하고 작업의 지연시간을 분석한다. 다른 프로세서에 할당된 작업들이 수행중인 자원을 요구할 때, 자원을 수행하는 작업의 우선순위를 높여줌으로써 자원수행을 빠르게 종료하게 한다. 이로 인해 자원에 의한 지연을 최소화한다. 특히, 높은 우선순위 작업의 경우 더욱 작은 지연시간을 갖게한다. 시뮬레이션을 통한 Shared Memory Protocol [5]과의 비교, 분석 결과 성능의 향상을 보임을 알 수 있다. 다양한 작업집합에 대한 지연시간을 분석하였다.

• 한국HCI학회:학술대회논문집
• /
• 2006.02a
• /
• pp.1175-1181
• /
• 2006

본 논문에서는 모바일 그리드 컴퓨팅 환경에서, 작업분배에 대한 보다 효율적인 계획을 스스로 결정하고, 이를 지속적으로 개선하는 'Self-Growing Engine' 기반 적응시스템을 제안한다. 최근, 모바일 컴퓨팅 환경에서의 다양한 제약사항을 극복하기 위해, 주변의 여러 컴퓨팅 단말기들의 유휴자원을 공유하여 하나의 작업을 처리하는 그리드 컴퓨팅 개념을 무선환경에 적용하려는 연구가 이슈로 등장하고 있다. 이때, 대부분의 기존 연구들은 그리드 컴퓨팅에 참여하는 단말기들의 리소스 상태만 고려하여 작업을 할당하는 방식을 취하고 있다. 따라서 상대적으로 작업효율이 낮은 단말기에서 작업이 할당되는 경우도 생기게 된다. 제안 시스템에서는 보다 효율적인 작업분배 결정을 위해, 다양한 사항을 고려하여 적절한 단말기를 선택하며, 각 작업수행결과를 history로 저장하며, 이후에 같은 요청이 있을 때 이를 분석하여 보다 적절한 단말기를 선택하도록 스스로 진화하는 특성을 갖는다. 우리는 제안 시스템의 평가를 위해, 데스크 탑에서 프로토타입을 구현하여 시뮬레이션을 수행하였으며, 그 결과를 통해 제안시스템의 효율성을 증명하였다.

• Journal of KIISE:Computer Systems and Theory
• /
• v.30 no.1
• /
• pp.1-11
• /
• 2003

Fragmentation is the main performance bottleneck of large, multi-user multicomputer system. This paper presents an L-Shaped Submesh Allocation(LSSA) strategy, which lifts the restriction on the rectangular shape formed by allocated processors in order to address the problem of fragmentation. LSSA can manipulate the shape of the required submesh to fit into the fragmented mesh system. Thus, LSSA accommodates incoming jobs faster than other strategies and results in the reduction of job response time. Extensive simulations show that LSSA performs more efficiently than other strategies in terms of the external fragmentation, the job response time and the system utilization.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.9 no.1
• /
• pp.165-171
• /
• 2005

In this paper, the centralized dynamic load-balancing was processed effectively by using the network based parallel genetic algorithm. Unlike the existing method using genetic algorithm, the performance of central scheduler was improved by distributing the process for the searching of the optimal task assignment to clients. A roulette wheel selection and an elite preservation strategy were used as selection operation to improve the convergence speed of optimal solution. A chromosome was encoded by using sliding window method. And a cyclic crossover was used as crossover operation. By the result of simulation for the performance estimation of central scheduler according to the change of flexibility of load-balancing method, it was verified that the performance is improved in the proposed method.

• The KIPS Transactions:PartA
• /
• v.12A no.4 s.94
• /
• pp.281-288
• /
• 2005

Efficient utilization of processing resources in a large multicomputer system with the possibility of fault occurrence depends on the reliable processor management scheme. This paper presents a dynamic and reliable processor allocation strategy to increase the performance of mesh-connected parallel systems with faulty processors The basic idea is to reconfigure a faulty mesh system into a maximum convex system using the fault-free upper or lower boundary nodes to compensate for the non-boundary faulty nodes. To utilize the non-rectangular shaped system parts, our strategy tries to allocate L-shaped submeshes instead of signaling the allocation failure. Extensive simulations show that the strategy performs more efficiently than other strategies in terms of the job response time md the system utilization.

• Journal of KIISE
• /
• v.44 no.5
• /
• pp.439-448
• /
• 2017

The workloads of large high-performance computing (HPC) scientific applications are steadily becoming "bursty" due to variable resource demands throughout their execution life-cycles. However, the over-provisioning of virtual resources for optimal performance during execution remains a key challenge in the scheduling of scientific HPC applications. While over-provisioning of virtual resources guarantees peak performance of scientific application in virtualized environments, it results in increased amounts of idle resources that are unavailable for use by other applications. Herein, we proposed a memory resource reconfiguration approach that allows the quick release of idle memory resources for new applications in OS-level virtualized systems, based on the applications resource-usage pattern profile data. We deployed a scientific workflow application in Docker, a light-weight OS-level virtualized system. In the proposed approach, memory allocation is fine-tuned to containers at each stage of the workflows execution life-cycle. Thus, overall memory resource utilization is improved.

• Journal of KIISE:Computer Systems and Theory
• /
• v.35 no.5
• /
• pp.187-198
• /
• 2008

This paper develops a framework for efficient resource management within the grid service environment. Resource management is the core element of the grid service; therefore, GridRMF(Grid Resource Management Framework) is modeled and developed in order to respond to such variable characteristics of resources as accordingly as possible. GridRMF uses the participation level of grid resource as a basis of its hierarchical management. This hierarchical management divides managing domains into two parts: VMS(Virtual Organization Management System) for virtual organization management and RMS(Resource Management System) for metadata management. VMS mediates resources according to optimal virtual organization selection mechanism, and responds to malfunctions of the virtual organization by LRM(Local Resource Manager) automatic recovery mechanism. RMS, on the other hand, responds to load balance and fault by applying resource status monitoring information into adaptive performance-based task allocation algorithm.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.16 no.1
• /
• pp.283-290
• /
• 2016

As the development of recent science and technology, high-performance computing resources is needed to solve complex problems. To reach these requirements, it has been actively studied about grid computing that consist of a huge system which bind a heterogeneous high performance computing resources into on which are geographically dispersed. However, The current research situation which are the process to obtain the best results in the limited resources and the scheduling policy to accurately predict the total execution time of the real-time task are very poor. In this paper, in order to overcome these problems, we suggested a grid computing-based risk management system which derived from the system structure and the process for improving the efficiency of the system, grid computing-based working methodology, risk policy module which can manage efficiently the problem of the work of resources(Agent), scheduling technique and allocation method which can re-allocate the resource allocation and the resources in problem, and monitoring which can manage resources(Agent).