• Progress in Medical Physics
• /
• v.2 no.1
• /
• pp.29-35
• /
• 1991

For accureate treatment planning, new models of dose calculations are being developed which require the knowledge of the energy spectra and angular distributions of the X-rays incident on the surface of the material. In this present study, we applied the Monte Carlo methods to the systematic analysis of the spectra distribution of X-ray beams produced by medical linear accelerator. As expected, the spectra become softer as the distance is farther from the central axis. also, its influenced by the geometrical dffect of head system.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.13 no.3
• /
• pp.255-259
• /
• 1988

The distributing current is calculated on the infinit plane mattaric grattings for the TM waves. The matrix is larger, when the moment method is applied this structure. So, the moment method of this case is required large memory and long CPU times. Those boundary condition and the scattering formura are transformed into spectal domain. Taking account of the peridic structure, this formular is changed in a series form by using the Flouquet mode. By making a suitable basis function, this equation is expreseed matrix form. So the distributing current on the mattaric strip is able to caculate by using this equation. We calculate magnitude of the distributing current for varing these spaces, widthes and an angle of incident waves.

• The KIPS Transactions:PartB
• /
• v.8B no.3
• /
• pp.231-236
• /
• 2001

일반적인 다층 신경망에서 가중치의 갱신 알고리즘으로 사용하는 오류 역전과 방식은 가중치 갱신 결과를 고정된(fixed) 한 개의 값으로 결정한다. 이는 여러 갱신의 가능성을 오직 한 개의 값으로 고정하기 때문에 다양한 가능성들을 모두 수용하지 못하는 면이 있다. 하지만 모든 가능성을 확률적 분포로 표현하는 갱신 알고리즘을 도입하면 이런 문제는 해결된다. 이러한 알고리즘을 사용한 베이지안 신경망 모형(Bayesian Neural Networks Models)은 주어진 입력값(Input)에 대해 블랙 박스(Black-Box)와같은 신경망 구조의 각 층(Layer)을 거친 출력값(Out put)을 계산한다. 이 때 주어진 입력 데이터에 대한 결과의 예측값은 사후분포(posterior distribution)의 기댓값(mean)에 의해 계산할 수 있다. 주어진 사전분포(prior distribution)와 학습데이터에 의한 우도함수(likelihood functions)에 의해 계산한 사후확률의 함수는 매우 복잡한 구조를 가짐으로 기댓값의 적분계산에 대한 어려움이 발생한다. 따라서 수치해석적인 방법보다는 확률적 추정에 의한 근사 방법인 몬테 칼로 시뮬레이션을 이용할 수 있다. 이러한 방법으로서 Hybrid Monte Carlo 알고리즘은 좋은 결과를 제공하여준다(Neal 1996). 본 논문에서는 Hybrid Monte Carlo 알고리즘을 적용한 신경망이 기존의 CHAID, CART 그리고 QUEST와 같은 여러 가지 분류 알고리즘에 비해서 우수한 결과를 제공하는 것을 나타내고 있다.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.34 no.6B
• /
• pp.568-574
• /
• 2009

320km optical transmission link for 128 channel DWDM (dense wavelength-division-multiplexing) signals is simulated and fabricated. An optical fiber amplifier for the link is composed of a distributed Raman amplifier and dual C/L-band EDFAs which are optimized for the performances of an optical amplifier obtained from the simulation. Gain and NF of the optimized EDFAs are above 19dB and below 7.5dB, respectively. The resultant OSNRs (optical signal to noise ratios) of the link are average 25dB on each band.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.25 no.5
• /
• pp.439-446
• /
• 2012

This paper presents an implicit moving least squares(MLS) difference method for improving the solution accuracy of 1-D free boundary problems, which implicitly updates the topology change of moving interface. The conventional MLS difference method explicitly updates the moving interface; it requires no iterative solution procedure but results in the loss of accuracy. However, the newly developed implicit scheme makes the total system nonlinear involving iterative solution procedure, but numerical verification show that it dramatically elevates the solution accuracy with moderate computation increase. Through numerical experiments for melting problems having moving singularity, it is verified that the proposed method can achieve the second order accuracy.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.11 no.8
• /
• pp.1297-1304
• /
• 2000

The full wave method requires a great number of computer memory and lasting long CPU time for the calculation of the discontinuity problems in microstrip structures. While the computation only for the transverse field components at those structures causes the both of time and memory reduction. For the case of the calculating only transverse components for the most of microstrip structures such as low-pass filter, branch coupler and patch antenna the computer memory and running time can be reduced to about 50% and 33%, comparing to the full wave computation. Consequently, the proposed method than that of TEM-mode has an advantages of higher speed and less memory than that of conventional FDTD analysis.

• 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 Mineralogical Society of Korea
• /
• v.31 no.1
• /
• pp.13-22
• /
• 2018

Aluminum silicates ($Al_2SiO_5$) undergo phase transitions among kyanite, andalusite, and sillimanite depending on temperature and pressure conditions. The minerals are often used as an important indicator of the degree of metamorphism for certain metamorphic rocks. In this study, we have applied classical molecular dynamics (MD) simulations and density functional theory (DFT) to the aluminum silicates. We examined the crystal structures as a function of applied pressure and the corresponding stabilities based on calculated enthalpies at each pressure. In terms of the lattice parameters, both methods showed that the volume decreases as the pressure increases as observed in the experiment. In particular, DFT results differed from experimental results by much less than 1%. As to the relative stability, however, both methods showed different levels of accuracy. In the MD simulations, a transition pressure at which the relative stability between two minerals reverse could not be determined because the enthalpies were insensitive to the applied pressure. On the other hand, in DFT calculations, the relative stability relation among the three minerals was consistent with experiment, although the transition pressure was strongly dependent on the choice of the electronic exchange-correlation functional.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.11 no.5
• /
• pp.17-30
• /
• 2001

In this paper, we implemented a scalar multiplier needed at an elliptic curve cryptosystem over standard basis in $GF(2^{163})$. The scalar multiplier consists of a radix-16 finite field serial multiplier and a finite field inverter with some control logics. The main contribution is to develop a new fast finite field inverter, which made it possible to avoid time consuming iterations of finite field multiplication. We used an algorithmic transformation technique to obtain a data-independent computational structure of the Extended Euclid GCD algorithm. The finite field multiplier and inverter shown in this paper have regular structure so that they can be easily extended to larger word size. Moreover they can achieve 100% throughput using the pipelining. Our new scalar multiplier is synthesized using Hyundai Electronics 0.6$\mu\textrm{m}$ CMOS library, and maximum operating frequency is estimated about 140MHz. The resulting data processing performance is 64Kbps, that is it takes 2.53ms to process a 163-bit data frame. We assure that this performance is enough to be used for digital signature, encryption & decryption and key exchange in real time embedded-processor environments.

• Journal of Broadcast Engineering
• /
• v.11 no.1 s.30
• /
• pp.138-141
• /
• 2006

This paper evaluates the computational efficiency of sample-rate converters with rational factors in multi-stage structure in terms of memory requirement and multiplications per second. We describe resolution preserving and mutual prime conditions, and then present a method for designing the converter from which optimal rational-valued conversion factors for each stage can be yielded directly. As an example, we show an implementation of the 44.1-to-48KHz sample-rate converter in 2-stage structure.