• The KIPS Transactions:PartA
• /
• v.12A no.2 s.92
• /
• pp.117-126
• /
• 2005

This paper presents an adaptive processor allocation strategy to reduce fragmentation in a large multi-user multicomputer system. A small number of jobs with unexpectedly large submesh allocation requirements may significantly increase the queuing delay of the rest of jobs. Under such circumstances, our strategy further tries to allocate L-shaped submeshes instead of signaling the allocation failure unlike other strategies. We have developed the efficient algorithm to find the allocatable L-shaped submeshes. Thus, our strategy reduces the mean response time by minimizing the queuing delay, even though jobs are scheduled in an FCFS to preserve fairness. The simulations show that our strategy performs more efficiently than other strategies in terms of the job response time and the system utilization.

• KIPS Transactions on Computer and Communication Systems
• /
• v.3 no.6
• /
• pp.173-178
• /
• 2014

By exploiting parallel processing, the proposed scheduling scheme enhances energy saving capability of multicore processors for real-time tasks while satisfying deadline and fault tolerance constraints. The scheme searches for a near minimum-energy schedule within a polynomial time, because finding the minimum-energy schedule on multicore processors is a NP-hard problem. The scheme consumes manifestly less energy than the state-of-the-arts method even with low parallel processing speedup as well as with high parallel processing speedup, and saves the energy consumption up to 86%.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.38 no.5
• /
• pp.16-26
• /
• 2001

In this paper, we present a new scheduling method called CDSS(Carried-Dependence Self-Scheduling) for efficiently execution of the loop with intra dependency between iterations based on the central queue. We also implemented it on shared memory system using Java language. Also, we study the modification that converts the existing self-scheduling method based on the central task queue for parallel loops onto the same form applied to loop with loop-carried dependences. The proposed method is self scheduling and assigns the loops in three-level considering the synchronization point according to the dependence distance of the loops. To adapt the proposed scheme and modified methods into various platforms, including a uni-processor system, we use threads for implementation. Compared to other assignment algorithms with various changes of application and system parameters, CDSS is found to be more efficient than other methods in overall execution time including scheduling overheads. CDSS shows improved performance over modified SS, Factoring, GSS and CSS by about 0.02, 40.5, 46.1 and 53.6%, respectively. In CDSS, we achieve the best performance on varying application programs using a few threads, which equal the dependence distance.

• The Transactions of the Korea Information Processing Society
• /
• v.7 no.1
• /
• pp.12-19
• /
• 2000

In the development of a production process scheduling system a collaboration method must be defined for the cooperation among submodules within the system. The blackboard architecture is exploited for solving the collaboration problem, which is one of the problem solving architecture that belongs to the distributed artificial intelligence. The dynamic states of the problem solving processes are presented in the hierarchically constructed shared working memory called as a blackboard. The communication for the collaboration is done through the blackboard. The problem solving steps are contained in the global controller, one of a component that consists the blackboard architecture, as knowledge. The global controller activates proper submodules based on the knowledge. By applying the blackboard architecture the collaboration problem among submodules in the grating production process scheduling system (GPSS) has been solved as well as the system became adaptable to the future modifications and expansions.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.18 no.3
• /
• pp.69-77
• /
• 2008

In the recent years, power attacks were widely investigated, and so various countermeasures have been proposed. In the case of block ciphers, masking methods that blind the intermediate results in the algorithm computations(encryption, decryption, and key-schedule) are well-known. Applications of masking methods are able to vary in different block ciphers, therefore suitable masking methods about each ciphers have been researched. Existed methods of ARIA have many revisions of mask value. And because existed masking methods pay no regard for key schedule, secret information can be exposed. In the case of ARIA, this problem is more serious than different block ciphers. Therefore we proposes an efficient masking scheme of ARIA including the key-schedule. Our method reduces time-complexity of ARIA encryption, and solve table-size problem of the general ARIA masking scheme from 256*8 byte to 256*6 byte.

• Journal of the Korea Society for Simulation
• /
• v.19 no.1
• /
• pp.1-11
• /
• 2010

K company produces semiconductor package products under the make-to-order policy to supply for domestic and foreign semiconductor manufacturing companies. Its production process is a machine-paced assembly line type, which consists of die sawing, assembly, and test. This paper suggests three plans to increase process throughput based on the process analysis of K company and evaluates them via a simulation model using a real data collected. The three plans are line balancing by adding machines to the bottleneck process, product group scheduling, and reallocation of the operators in non-bottleneck processes. The evaluation result shows the highest daily throughput increase of 17.3% with an effect of 2.8% reduction of due date violation when the three plans are applied together. Payback period for the mixed application of the three plans is obtained as 1.37 years.

• The Transactions of the Korea Information Processing Society
• /
• v.6 no.11
• /
• pp.3124-3130
• /
• 1999

A duplex system enhances reliability by tolerating faults through spatial redundancy. Faults can be detected by duplicating identical tasks in pairs of modules. However, this kind of systems cannot even detect the fault if it occurs coincidently due to either malfunctions of common component such as power supply and clock or due to such environmental disruption as EMI. In the paper, we propose a method to reduce those effects of coincident faults in the duplex controller computer. Specifically, a duplex system tolerates coincident faults by using a sophistication sequencing of scheduling technique with certain timing redundancy. In particular when all tasks should be completed in the sense of real-time, the suggested scheduling method works properly to minimize the probability of faulty tasks due to coincident fault without missing the timing constraints.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2006.05a
• /
• pp.315-318
• /
• 2006

본 논문에서는 TV OSD(On Screen Display) 메뉴 자동검증 시스템에서 문자패턴의 실시간 인식을 위한 방법론을 고찰한다. 이는 일반적인 문자인식 문제와는 달리 시스템 환경에 대한 몇 가지 가정과 제약조건을 고려해야 한다. 예컨대 문제의 특성상 카메라 및 TV 제어 기기부의 동작과 연동하는 작업 스케쥴링 기능과 실시간 분석기능 등의 요건은 시스템개발을 복잡하게 하는 반면, 주어진 OSD 메뉴 데이터로부터 검증과정은 미지 패턴에 대한 인식과정을 단순화하여 일종의 판정(decision) 문제로 고려될 수 있게 한다. 본 연구에서는 인식의 방법론으로서 수정된 구조의 FMM 신경망을 적용한다. 이는 하이퍼박스 기반의 패턴 분류기로서 간결하면서도 강력한 학습기능을 제공한다. 기존의 FMM 모델이 갖는 단점인 학습패턴에서 특징분포와 빈도를 고려하지 못한다는 점을 개선하여, 특징과 하이퍼박스간의 가중치 요소를 고려한 활성화 특성을 정의한다. 또한 실제 데이터를 사용한 실험결과를 통해 제안된 이론의 유용성을 고찰한다.

• Journal of Korean Institute of Industrial Engineers
• /
• v.25 no.3
• /
• pp.360-368
• /
• 1999

We consider the problem of scheduling n jobs with sequence-dependent processing times on a set of parallel-identical machines. The processing time of each job consists of a pure processing time and a sequence-dependent setup time. The objective is to maximize the total remaining machine available time which can be used for other tasks. For the problem, a hybrid genetic algorithm is proposed. The algorithm combines a genetic algorithm for global search and a heuristic for local optimization to improve the speed of evolution convergence. The genetic operators are developed such that parallel machines can be handled in an efficient and effective way. For local optimization, the adjacent pairwise interchange method is used. The proposed hybrid genetic algorithm is compared with two heuristics, the nearest setup time method and the maximum penalty method. Computational results for a series of randomly generated problems demonstrate that the proposed algorithm outperforms the two heuristics.

• Journal of Korean Institute of Industrial Engineers
• /
• v.21 no.1
• /
• pp.3-16
• /
• 1995

JIT manufacturing requires that materials and services should be done neither too early nor tardy. Therefore, related jobs have to be performed just-in-time. In this paper, we consider a problem of determining a schedule of jobs with a common due date. The objective is to minimize the squared deviation of complete times from the due date. An artificial neural network which is based on the mean field theory is proposed for the problem. Computational results obtained from computer simulations show that the network performs very well.