• Journal of KIISE:Computer Systems and Theory
• /
• v.28 no.12
• /
• pp.632-641
• /
• 2001

In this paper, we introduce new methods for hiding computation overheads involved in load redistributing for parallel computer of hypercube, mesh and tree topologies. The basic idea is either coalescing some phases of load redistributing to overlap the transfer on different links or dividing each phase into steps to pipeline the transfer of load unit by unit for maximum utilization of links. They proved effective in making links busy transmitting load as soon as possible, hence reducing the computation overheads involved in balancing. Proposed techniques experimented on hypercube, mesh or tree topologies reduce communication overheads by 20% to 50% compared with known methods.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.8
• /
• pp.1930-1939
• /
• 1997

In this paper, we analyze the fault diameter and fault tolerance of Gray cube proposed recently in [12]. fault diameter of an interconnection network is one of the important network measures concerning the distance between nodes when some nodes fail. It is showed that fault diameter of n-dimensional Gray cube having $2^n$ nodes is [(n+1)/2]+2, ($n{\ge}3$). It means the increment of the longest distance between nodes under node-failure is only constant factor. Comparing the result with the fault diameter of well-known hypercube, the longest routing distance of a message in a Gray cube under node-failure is about the half of that hypercube.

• Journal of the Institute of Convergence Signal Processing
• /
• v.5 no.3
• /
• pp.236-242
• /
• 2004

In this paper, we propose an approach to parallelize a load balancing algorithm that was shown to be very effective in distributing workload for parallel computations. Load balancing algorithms are required in executing parallel program efficiently As a parallel computation model, we used dynamically growing tree structure that can be found in many application problems. The load balancing algorithm tries to balance the workload among processors while keeping the communication cost under certain limit. We show how the load balancing algorithm is effectively parallelized on mesh and hypercube interconnection networks, and analyzed the time complexity for each case to show that parallel algorithm actually reduced the various overhead.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.22 no.6
• /
• pp.763-772
• /
• 2019

There is a need to use sheath flow nozzle to detect bioaerosol such as virus and bacteria due to their characteristics. In order to enhance the detection performance depending on nozzle parameters, numerical analysis was carried out using a commercial code, ANSYS CFX. Eulerian-lagrangian approach method is used in this simulation. Multiphase flow characteristics between primary fluid and solid were considered. The detection performance was evaluated based on the results of flow field in nozzle chamber such as focusing efficiency and swirl strength. In addition, Latin hypercube sampling(LHS) of design of experiment(DOE) was used for generating a near-random sampling. Then, the acceleration section is optimized using response surface method(RSM). Results show that the optimized model achieved a 6.13 % in a focusing efficiency and 11.47 % increase in swirl strength over the reference model.

• Journal of the military operations research society of Korea
• /
• v.33 no.1
• /
• pp.59-74
• /
• 2007

Computer simulation models are in general quite complex and time-consuming to run, and therefore, a simpler meta-model is usually constructed for further analysis. In this paper, JANUS, a war-game simulator, is used to describe a certain tank combat situation. Then, second-order response surface and artificial neural network meta-models are developed using the data from eight different experimental designs. Relative performances of the developed meta-models are compared in terms of the mean squared error of prediction. Computational results indicate that, for the given problem, the second-order response surface meta-model generally performs better than the neural network, and the orthogonal array-based Latin hypercube design(LHD) or LHD using maximin distance criterion may be recommended.

• The KIPS Transactions:PartA
• /
• v.15A no.6
• /
• pp.345-350
• /
• 2008

Recently, a HyperStar network HS(2n,n) has been introduced as a new interconnection network of new topology for parallel processing. HyperStar network has properties of hypercube and star graph, further improve the network cost of a hypercube with the same number of nodes. In this paper, we show a construction algorithm of edge-disjoint spanning trees in HyperStar network HS(2n,n). Also, we prove that edge-disjoint spanning tree by the algorithm is optimal.

• Nuclear Engineering and Technology
• /
• v.42 no.6
• /
• pp.690-703
• /
• 2010

Sampling based uncertainty analysis was carried out to quantify uncertainty in predictions of best estimate code RELAP5/MOD3.2 for a thermal hydraulic test (10% hot leg break LOCA) performed in the Large Scale Test Facility (LSTF) as a part of an IAEA coordinated research project. The nodalisation of the test facility was qualified for both steady state and transient level by systematically applying the procedures led by uncertainty methodology based on accuracy extrapolation (UMAE); uncertainty analysis was carried out using the Latin hypercube sampling (LHS) method to evaluate uncertainty for ten input parameters. Sixteen output parameters were selected for uncertainty evaluation and uncertainty band between $5^{th}$ and $95^{th}$ percentile of the output parameters were evaluated. It was observed that the uncertainty band for the primary pressure during two phase blowdown is larger than that of the remaining period. Similarly, a larger uncertainty band is observed relating to accumulator injection flow during reflood phase. Importance analysis was also carried out and standard rank regression coefficients were computed to quantify the effect of each individual input parameter on output parameters. It was observed that the break discharge coefficient is the most important uncertain parameter relating to the prediction of all the primary side parameters and that the steam generator (SG) relief pressure setting is the most important parameter in predicting the SG secondary pressure.

• Journal of the Korean Society of Industry Convergence
• /
• v.22 no.4
• /
• pp.387-394
• /
• 2019

This paper presents an analytical study on the strength and stiffness of various types of truss structures. The applied models are triangular-like opened truss-wall triangular model (OTT), closed truss-wall triangular model (CTT), opened solid-wall triangular model (OST), and hypercube models defined as core-filled or core-spaced cube. The models are analyzed by numerical model analysis using DEFORM 2D/3D tool with AISI 304 stainless steel. Then, the ideal solutions for stiffness and strength are defined. Finally, the relative elastic modulus of the core-spaced model is obtained as 0.0009, which is correlated with the cancellous bone for the relative density range of 0.029-0.03, and the relative elastic modulus for the core-filled model is obtained as 0.0015, which is correlated with cancellous bone for the relative density range of 0.035-0.036. For the relative compressive yield strength, the OTT reasonably agrees with the cancellous bone for the relative density of 0.042 and the relative compressive strength of 0.05. The CTT and OST are in good agreement at the relative density of 0.013 and the relative compressive yield strength of 0.002. The hypercube models can be used for the cancellous bone for stiffness, and the triangular models can be used for the cancellous bone for strength. However, none of the models can be used to replace the compact bone because it requires much higher stiffness and strength. In the near future, compact bone replacement must be further studied. In addition, previously mentioned models should be developed further.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.22 no.4
• /
• pp.527-535
• /
• 2019

Hanger in a rail-launched missile protrudes in general and causes to increase significant drag force. One method to avoid the significant increase of drag force is to apply fairings on the hanger. In this paper, sloping shaped fairing parameters of height, width, and length are optimized to minimize the drag force under subsonic speed region by examining three configurations of fairings : front-fairing only, rear-faring only, and the both front and rear fairing. We use Latin Hypercube Sampling method to determine the experimental points, and computational fluid dynamics with incompressible RANS solver was applied to acquire the data at sampling points. Then, we construct a meta model by kriging method. We find the best choice among three configurations examined : both front and rear fairing reduce the drag force by 63 % without the constraint of fairing mass, and front fairing reduced the drag force by 52 % with the constraint of hanger mass.

• Nuclear Engineering and Technology
• /
• v.26 no.1
• /
• pp.78-89
• /
• 1994

The radionuclide release rates from the engineered barrier composed of concrete structure and clay-based backfill in a low and intermediate level waste repository were assessed. Four types of release pathway were considered, and the contribution of each pathway to the total release were analyzed. To quantify the effect of uncertainties of input parameter values on the assessment of radionuclide release rates, the Latin Hypercube sampling method was used, and the resulting release rate distribution were determined through a goodness-of-fit test. Finally, the ranges of maxi-mum release rates ore estimated statistically with a confidence level of 95%.