• Journal of Communications and Networks
• /
• v.12 no.6
• /
• pp.600-604
• /
• 2010

A method of width adaptation in the radial basis function network (RBFN) using stochastic gradient (SG) algorithm is introduced. Using Taylor's expansion of error signal and differentiating the error with respect to the step-size, the optimal time-varying step-size of the width in RBFN is derived. The proposed approach to adjusting widths in RBFN achieves superior learning speed and the steady-state mean square error (MSE) performance in nonlinear channel environment. The proposed method has shown enhanced steady-state MSE performance by more than 3 dB in both nonlinear channel environments. The results confirm that controlling over step-size of the width in RBFN by the proposed algorithm can be an effective approach to enhancement of convergence speed and the steady-state value of MSE.

• International Journal of Aeronautical and Space Sciences
• /
• v.5 no.2
• /
• pp.43-53
• /
• 2004

Collision avoidance for multiple aircraft can be stated as a problem ofmaintaining safe distance between aircraft in conflict. Optimal collision avoidanceproblem seeks to minimize the given cost function while simultaneously satisfyingconstraints. The cost function could be a function of time or control input. This paper addresses the trajectory time-optimization problem for collision avoidance of unmanned aerial vehicles(UAVs). The problem is difficult to handle in general due to the two-point boundary value problem subject to dynamic environments. Some simplifying aleorithms are used for potential applications in on-line operation.Although under possibility of more complicated problems, a dynamic problem is transformed into a static one by prediction of the conflict time and some appropriate assumptions.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.29B no.3
• /
• pp.16-24
• /
• 1992

In this paper, we solve the problem of designing a robust optimal controller which minimizes the H$\infty$-norm of the mixed sensitivity function matrix for linear multivariable systems. For a given minimized value, ${\gamma}$>o, an algorithm of finding all stabilizing controllers, such that the H$\infty$-norm of the mixed sensitivity function matrix is less than ${\gamma}$, is developed. The proposed algorithm, which is based on the model-matching and the interpolation theory, can be used for the H$\infty$-optimization problem.

• Journal of Korean Society of Steel Construction
• /
• v.17 no.1 s.74
• /
• pp.13-21
• /
• 2005

Advanced analysis and optimal design of steel arch bridges is presented. In the design method using an advanced analysis, separate member capacity checks after analysis are not required because the stability and strength of the structural system and its component members can be rigorously treated in the analysis. The geometric nonlinearity is considered by using the stability function. The Column Research Council tangent modulus is used to account for gradual yielding due to residual stresses. A parabolic function is used to represent the transition from elastic to zero stiffness associated with a developing hinge. An optimization technique used is a modified section increment method. The member with the largest unit value evaluated by AASHTO-LRFD interaction equation is replaced one by one with an adjacent larger member selected in the database. The objective function is taken as the weight of the steel arch bridge and the constraint functions account for load-carrying capacities and deflection requirements. Member sizes determined by the proposed method are compared with those given by other approaches.

• The Journal of the Acoustical Society of Korea
• /
• v.40 no.4
• /
• pp.347-358
• /
• 2021

The objective of this study is to optimize the shape of doubly curved surface where a conformal array is equipped. That surface is modeled with a double-ellipsoid solid controlled by four parameters. By analyzing the performance of the conformal array beams with the beam pattern synthesis, two design parameters are determined. Then, we define the weighted object function which is formulated as the sum of sharp indexes for directivity index, the elevation resolution, and the bearing resolution. The direct calculation on all grids is used to evaluate the weighted object function and seek the optimal value of two design parameters when the weightings are given. In the simulation, four kinds of weighting cases are respectively applied to evaluate the weighted object function. The optimal shapes of double-ellipsoid solid are shown in each case. Especially, when the uniform weightings are used, the double-ellipsoid solid with more smooth surface is obtained.

• Health Policy and Management
• /
• v.4 no.1
• /
• pp.107-122
• /
• 1994

Many alternatives have been discussed to reduce the medical expenditure and to use the medical resources effectively. Many studies about the economies of scale have been done for the last several decades. This study has analyzed the relationship between the number of beds and the mean expense per hospitalization day in Korea. A Cost Function Model was identified and we wanted to see the minimum optimal size with the cheapest mean expense per hospitalization day. The result is as follows; 1. In the Cost Function Mode, (the number of beds)$^{2}$, the number of personnel, productivity and training institutions are the factors that statisticaly influence the mean expenses. 2. By the univariate analysis the mean expense proved to be the smallest as the level of 150-200bed, The breaked down of the components of expenses shows that the mean labor cost is much different from the mean value of material and administration costs, and that hospital with 150-200 beds also have the minimal expense. The mean expense goes up dramatically in hospitals of 450 beds or more. 3. When the other conditions are constant, according to the multiple regression analysis of the mean expense per adjusted hospitalization day the minimum optimal size with the cheapest expense is a hospital with 191 beds and the hospital with 230 beds takes the lowest mean labor cost. The material or administration costs are not influenced by hospital size. This research has limitation in measuring the variables that influence hospital xpenses, in estimating hospital output by the number of beds in considering outpatient cost and in securing representativeness of hospitals because many hospitals made no responses to the research questionnare. But it is valuable and helpful for development of health policy to figure out the number of beds with the cheapest expense per hospitalization day.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.19 no.40
• /
• pp.233-240
• /
• 1996

A fixed amount compensator is proposed for a process with a linear tool wear function. A Cost model is constructed which involve process adjustment cost and quality loss. Symmetric and asymmetric quadratic functions of the deviation of a quality measurement from the nominal target value are considered as the quality loss functions. Methods of finding optimal values of initial setting and compensation limit are presented and a numerical example is given.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.7 no.1
• /
• pp.11-16
• /
• 1982

A method is described for the solution of a linearly constrained program with parametric nonlinear objective function. The algorithm proposed in this paper may be regarded as an extension of the simplex method for parametric linear programming. Namely, it specifies the basis at each stage such that feasibility ana optimality of the original problem are satisfied by the optimal solution of the reduced parametric problem involving only nonbasic variables. It is shown that under appropriate assumptions the algorithm is finite. Parametric procedures are also indicated for solving each reduced parametric problem by maintaining the Kuhn-Tucker conditions as the parameter value varies.

• 제어로봇시스템학회:학술대회논문집
• /
• 1989.10a
• /
• pp.884-889
• /
• 1989

In this paper, a new procedure for selecting weighting matrices in linear discrete time quadratic optimal control problem (LQ-problem) is proposed. In LQ-problems, the quadratic weighting matrices are usually decided on trial and error in order to get a good response. But using the proposed method, the quadratic weights are decided in such a way that all poles of the closed loop system are located in a desired region for good responses as well as for stability and values of the quadratic cost function are kept less then a specified value.

• 제어로봇시스템학회:학술대회논문집
• /
• 1992.10b
• /
• pp.208-213
• /
• 1992

In this paper we are concerned with optimal control problems whose costs am quadratic and whose states are governed by linear delay equations and general boundary conditions. The basic new idea of this paper is to Introduce a new class of linear operators in such a way that the state equation subject to a starting function can be viewed as an inhomogeneous boundary value problem in the new linear operator equation. In this way we avoid the usual semigroup theory treatment to the problem and use only linear operator theory.