• The Pure and Applied Mathematics
• /
• v.15 no.1
• /
• pp.47-55
• /
• 2008

Local convergence results for three Newton-like methods in Banach space are provided. A comparison is given between the three convergence radii. Then we show that using the largest convergence radius we can pick an initial guess from with we start the corresponding iteration. It turns out that after a finite number of steps we can always use the iterate found as the starting guess for a faster method, since this iterate will be inside the convergence domain of the new method.

• The Journal of Information Technology
• /
• v.8 no.1
• /
• pp.119-124
• /
• 2005

Learning procedure in the neural network is updating of weights between neurons. Unadequate initial learning coefficient causes excessive iterations of learning process or incorrect learning results and degrades learning efficiency. In this paper, adaptive learning algorithm is proposed to increase the efficient in the learning algorithms of Kohonens Self-Organization Neural networks. The algorithm updates the weights adaptively when learning procedure runs. To prove the efficiency the algorithm is experimented to clustering of the random weight. The result shows improved learning rate about 42~55% ; less iteration counts with correct answer.

• Journal of the Chungcheong Mathematical Society
• /
• v.16 no.1
• /
• pp.113-121
• /
• 2003

This paper is concerned with some properties of stable domains and limit functions. Using the relationship between cycles of periodic stable domains and orbits of critical points and using the Sullivan theorem [19], we prove that the value of a constant limit function in some stable domain for a rational function f of degree at least two lies in the closure of the set of critical orbits of f.

• Journal of the Chungcheong Mathematical Society
• /
• v.8 no.1
• /
• pp.1-10
• /
• 1995

For analyzing systems of multi-variate nonlinear equations, the linearized modelling techniques are elaborated. The technique applies Newton-Raphson iteration, partial differentiation and matrix operation providing solvable solutions to complicated problems. Practical application examples are given in; determining the zero point of functions, determining maximum (or minimum) point of functions, nonlinear regression analysis, and solving complex co-efficient polynomials. Merits and demerits of linearized modelling techniques are also discussed.

• Journal of the Chungcheong Mathematical Society
• /
• v.15 no.2
• /
• pp.47-56
• /
• 2003

This note is concerned with some properties of fixed points and periodic points. First, we have constructed a generalized continuous function to give a proof for the fact that, as the reverse of the Sharkovsky theorem[16], for a given positive integer n, there exists a continuous function with a period-n point but no period-m points wherem is a predecessor of n in the Sharkovsky ordering. Also we show that the composition of two transcendental meromorphic functions, one of which has at least three poles, has infinitely many fixed points.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• v.15 no.1
• /
• pp.19-30
• /
• 2011

In [21], we compared the Newton-Krylov method and some high-order methods to solve nonlinear systems. In this paper, we propose high-order Newton-Krylov methods combining the Newton-Krylov method with some high-order iterative methods to solve systems of nonlinear equations. We provide some numerical experiments including comparisons of CPU time and iteration numbers of the proposed high-order Newton-Krylov methods for several nonlinear systems.

• Korean Journal of Mathematics
• /
• v.19 no.2
• /
• pp.171-180
• /
• 2011

In this paper, we study the superstability problem bounded by two-variables of Th. M. Rassias type for the generalized sine functional equations $$g(x+y)f(x-y)=f(x)^2-f(y)^2 \\ f(x+y)g(x-y)=f(x)^2-f(y)^2 \\ g(x+y)g(x-y)=f(x)^2-f(y)^2$$, which does not use his iteration method.

• Korean Journal of Computational Design and Engineering
• /
• v.2 no.4
• /
• pp.237-243
• /
• 1997

A new method is introduced for the interpolation in NURBS Surface. This method uses the basis functions to assign the parameter values to the arbitrary set of geometric data and uses the iteration method to compute the control net. The advantages of this method are the feasible transformation of the data set to the matrix form and the effective surface generation as a result, especially to the design engineer.

• Journal of the Korean Mathematical Society
• /
• v.39 no.2
• /
• pp.319-330
• /
• 2002

The method of quasilinearization generates a monotone iteration scheme whose iterates converge quadratically to a unique solution of the problem at hand. In this paper, we apply the method to two families of three-point boundary value problems for second order ordinary differential equations: Linear boundary conditions and nonlinear boundary conditions are addressed independently. For linear boundary conditions, an appropriate Green\`s function is constructed. Fer nonlinear boundary conditions, we show that these nonlinearities can be addressed similarly to the nonlinearities in the differential equation.

• ETRI Journal
• /
• v.25 no.1
• /
• pp.1-8
• /
• 2003

This paper introduces an efficient iterative decoding method for high-dimensional block turbo codes. To improve the decoding performance, we modified the soft decision Viterbi decoding algorithm, which is a trellis-based method. The iteration number can be significantly reduced in the soft output decoding process by applying multiple usage of extrinsic reliability information from all available axes and appropriately normalizing them. Our simulation results reveal that the proposed decoding process needs only about 30% of the iterations required to obtain the same performance with the conventional method at a bit error rate range of $10^{-5}\;to\;10^{-6}$.