• Proceedings of the Korea Electromagnetic Engineering Society Conference
• /
• 2005.11a
• /
• pp.169-174
• /
• 2005

In this paper, we proposed a new algorithm of the inverse scattering for the reconstruction of unknown dielectric scatterers using the finite-difference time-domain method and the design sensitivity analysis. We introduced the design sensitivity analysis based on the gradient for the fast convergence of the reconstruction. By introducing the adjoint variable method for the efficient calculation, we derived the adjoint variable equation. As an optimal algorithm we used the steepest descent method and reconstructed the dielectric targets using the iterative estimation. To verify our algorithm we will show the numerical examples for the two-dimensional $TM^2$ cases.

• Proceedings of the Korean Society of Computational and Applied Mathematics Conference
• /
• 2003.09a
• /
• pp.9-9
• /
• 2003

Let $\Omega$ be a bounded, open, and polygonal domain in $R^2$ with re-entrant corners. We consider the following Partial Differential Equations: $$(I-\nabla\nabla\cdot+\nabla^{\bot}\nabla\times)u\;=\;f\;in\;\Omega$$, $$n\cdotu\;0\;0\;on\;{\Gamma}_{N}$$, $${\nabla}{\times}u\;=\;0\;on\;{\Gamma}_{N}$$, $$\tau{\cdot}u\;=\;0\;on\;{\Gamma}_{D}$$, $$\nabla{\cdot}u\;=\;0\;on\;{\Gamma}_{D}$$ where the symbol $\nabla\cdot$ and $\nabla$ stand for the divergence and gradient operators, respectively; $f{\in}L^2(\Omega)^2$ is a given vector function, $\partial\Omega=\Gamma_{D}\cup\Gamma_{N}$ is the partition of the boundary of $\Omega$; nis the outward unit vector normal to the boundary and $\tau$represents the unit vector tangent to the boundary oriented counterclockwise. For simplicity, assume that both $\Gamma_{D}$ and $\Gamma_{N}$ are nonempty. Denote the curl operator in $R^2$ by $$\nabla\times\;=\;(-{\partial}_2,{\partial}_1$$ and its formal adjoint by $${\nabla}^{\bot}\;=\;({-{\partial}_1}^{{\partial}_2}$$ Consider a weak formulation(WF): Find $u\;\in\;V$ such that $$a(u,v):=(u,v)+(\nabla{\cdot}u,\nabla{\cdot}v)+(\nabla{\times}u,\nabla{\times}V)=(f,v),\;A\;v{\in}V$$. (2) We assume there is only one singular corner. There are many methods to deal with the domain singularities. We introduce them shortly and we suggest a new Finite Element Methods by using Singular representation for the solution.

• Journal of the Korean Society of Fisheries and Ocean Technology
• /
• v.33 no.1
• /
• pp.68-75
• /
• 1997

In this paper, we propose the improved initial image estimation method for a fast fractal image decoding. When the correlation between a domain and a range is given as the linear equation, the value of initial image estimation using the conventional method is the intersection between its linear equation and y=x. If the gradient of linear equation is large, that the difference of the value between each adjacent pixels is large, the conventional method has disadvantage which has the impossibility of exact estimation. The method of the proposed initial image estimation performs well by two steps. he first step can improve the disadvantage of the conventional method. The second step upgrades the range value which was found previous step by referring information of its domain. Though the computational complexity for the initial image estimation increses slightly, the total computational complexity decreases by 30% than that of the conventional method because of diminishing in the number of iterations.

• Journal of Environmental Science International
• /
• v.1 no.2
• /
• pp.15-23
• /
• 1992

Development of cumulus is studied by numerically integrating the equation of motion equations of conservation for water vapor mixing ratio, and the thermodynamic energy equuation. We use the terrain-following coordinate system called z'-coordinate system, in which we can easily treat any calculation domain with terrain configuration such as mountains. The model domain of calculation is restricted vertically to 4.Skin and horizontally to 100 km, has a bell-type mountain in the centeral part. Four cases are considered, one in a neutral environment, second in a slightly stable environment, third in a environment decreasing water content with low value of initial water vapor mixing ratio, the fourth in a case with higher vapor gradient. The more the atmosphere is unstable, the more cumulus develops easily and the more water vapors is abundant, the more cumulus develops easily too. More detailed cloud microphysics parameterizations and wet deposition must be conridered to use in air pollutants prediction model.

• Proceedings of the Korean Fiber Society Conference
• /
• 2003.04a
• /
• pp.191-194
• /
• 2003

In this study, permeability tensor is computed for a three dimensional braided preform by applying a FVM to a unit cell. It is evident that resin will flow through the inter-tow and the intra-tow regions when pressure gradient is applied to the resin. The intra-tow region in the braided preform is regarded as not only an impermeable solid but also a permeable porous media. When the intra-tow region is excluded from domain of flow analysis in the case of the impermeable solid, the Stokes equation is computer for only inter-tow region. (omitted)

• Proceedings of the IEEK Conference
• /
• 1998.10a
• /
• pp.1273-1276
• /
• 1998

The HMM-Net is an architecture for a neural network that implements a hidden Markov model (HMM). The architecture is developed for the purpose of combining the discriminant power of neural networks with the time-domain modeling capability of HMMs. Criteria used for learning HMM-Net classifiers are maximum likelihood (ML), maximum mutual information (MMI), and minimization of mean squared error(MMSE). In this paper we propose an efficient learning method of HMM-Net classifiers using hybrid criteria, ML/MMSE and MMI/MMSE, and report the results of an experimental study comparing the performance of HMM-Net classifiers trained by the gradient descent algorithm with the above criteria. Experimental results for the isolated numeric digits from /0/ to /9/ show that the performance of the proposed method is better than the others in the respects of learning and recognition rates.

• Structural Engineering and Mechanics
• /
• v.58 no.2
• /
• pp.217-230
• /
• 2016

Effects of temperature dependent material properties on mixed mode fracture parameters of functionally graded materials subjected to thermal loading are investigated. A domain form of the $J_k$-integral method including temperature-dependent material properties and its numerical implementation using finite element analysis is presented. Temperature and displacement fields are calculated using finite element analysis and are used to compute mixed mode stress intensity factors using the $J_k$-integral. Numerical results indicate that temperature-dependency of material properties has considerable effect on the mixed-mode stress intensity factors of cracked functionally graded structures.

• Journal of the military operations research society of Korea
• /
• v.17 no.1
• /
• pp.146-158
• /
• 1991

In this paper, we deal with the TSP (Traveling Salesperson Problem) which is well-known as NP-complete optimization problem. the TSP is applicable to network routing. task allocation or scheduling. and VLSI wiring. Well known numerical methods such as Newton's Metheod. Gradient Method, Simplex Method can not be applicable to find Global Solution but the just give Local Minimum. Exhaustive search over all cyclic paths requires 1/2 (n-1) ! paths, so there is no computer to solve more than 15-cities. Heuristic algorithm. Simulated Annealing, Artificial Neural Net method can be used to get reasonable near-optimum with polynomial execution time on problem size. Therefore, we are able to select the fittest one according to the environment of problem domain. Three methods are simulated about symmetric TSP with 30 and 50-city samples and are compared by means of the quality of solution and the running time.

• Proceedings of the IEEK Conference
• /
• 1998.06a
• /
• pp.933-935
• /
• 1998

The HMM-Net is an architecture for a neural network that implements a hidden markov model (HMM). The architecture is developed for the purpose of combining the discriminant power of neural networks with the time-domain modeling capability of HMMs. Criteria used for learning HMM-Net classifiers are maximum likelihood(ML) and minimization of mean squared error(MMSE). In this paper we propose an efficient learning method of HMM_Net classifiers using a ML-MMSE hybrid criterion and report the results of an experimental study comparing the performance of HMM_Net classifiers trained by the gradient descent algorithm with the above criteria. Experimental results for the isolated numeric digits from /0/ to /9/ show that the performance of the proposed method is better than the others in the repects of learning and recognition rates.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.39 no.6
• /
• pp.616-623
• /
• 1990

This paper includes the analysis and formulation of a new model for studying scattering from wire meshes that is more efficient and simpler to apply than the previous methods. In the new method, the conjugate gradient method is employed to improve each previous iterative and the fast Fourier transform (FFT) technique is utilized. A numerical computation of mesh scattering algorithm has been carried out in the Spectral Domain. A study on the electromagnetic properties such as reflection coefficients, induced currents and aperture fields has been presented and compared with data calculated by other methods to support the validity of the algorithm.