• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2005.04a
• /
• pp.301-304
• /
• 2005

In this paper, we propose a new fuzzy set-based polynomial neuron (FSPN) involving the information granule, and new fuzzy-neural networks - Fuzzy Set based Polynomial Neural Networks (FSPNN). We have developed a design methodology (genetic optimization using Genetic Algorithms) to find the optimal structure for fuzzy-neural networks that expanded from Group Method of Data Handling (GMDH). It is the number of input variables, the order of the polynomial, the number of membership functions, and a collection of the specific subset of input variables that are the parameters of FSPNN fixed by aid of genetic optimization that has search capability to find the optimal solution on the solution space. We have been interested in the architecture of fuzzy rules that mimic the real world, namely sub-model (node) composing the fuzzy-neural networks. We adopt fuzzy set-based fuzzy rules as substitute for fuzzy relation-based fuzzy rules and apply the concept of Information Granulation to the proposed fuzzy set-based rules.

• Journal of KIISE:Computer Systems and Theory
• /
• v.36 no.1
• /
• pp.21-25
• /
• 2009

In this paper, we study the problem to fit a piecewise-quadratic polynomial curve to points in the plane. The curve consists of quadratic polynomial segments and two points are connected by a segment. But it passes through a subset of points, and for the points not to be passed, the error between the curve and the points is estimated in $L^{\infty}$ metric. We consider two optimization problems for the above problem. One is to reduce the number of segments of the curve, given the allowed error, and the other is to reduce the error between the curve and the points, while the curve has the number of segments less than or equal to the given integer. For the number n of given points, we propose $O(n^2)$ algorithm for the former problem and $O(n^3)$ algorithm for the latter.

• East Asian mathematical journal
• /
• v.29 no.3
• /
• pp.279-292
• /
• 2013

It is well known that each kernel function defines a primal-dual interior-point method(IPM). Most of polynomial-time interior-point algorithms for linear optimization(LO) are based on the logarithmic kernel function([2, 11]). In this paper we define a new eligible kernel function and propose a new search direction and proximity function based on this function for LO problems. We show that the new algorithm has ${\mathcal{O}}((log\;p){\sqrt{n}}\;log\;n\;log\;{\frac{n}{\epsilon}})$ and ${\mathcal{O}}((q\;log\;p)^{\frac{3}{2}}{\sqrt{n}}\;log\;{\frac{n}{\epsilon}})$ iteration bound for large- and small-update methods, respectively. These are currently the best known complexity results.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.32 no.11
• /
• pp.924-929
• /
• 2008

The rate of wear of cam followers in a valve train system is mainly a function of contact stress between the cam and the follower, sliding velocity and hydrodynamic film thickness between the two mating surfaces. The wear or surface fatigue can be reduced by maximizing the elastohydrodynamic film thickness. In this paper, an attempt has been made to estimate the optimal specific film thickness of cam-follower system quantitatively. A general TES polynomial function with real values of exponents is developed and genetic algorithm (GA) is used as optimization techniques for maximizing the minimum specific film thickness. The optimization programs enumerate values of the exponents for synthesis of cam displacement curves. The results show that the minimum film thickness can be increased considerably, e.g. approximately 7% in this paper.

• Honam Mathematical Journal
• /
• v.35 no.2
• /
• pp.235-249
• /
• 2013

It is well known that each kernel function defines primal-dual interior-point method (IPM). Most of polynomial-time interior-point algorithms for linear optimization (LO) are based on the logarithmic kernel function ([9]). In this paper we define new eligible kernel function and propose a new search direction and proximity function based on this function for LO problems. We show that the new algorithm has $\mathcal{O}(({\log}\;p)^{\frac{5}{2}}\sqrt{n}{\log}\;n\;{\log}\frac{n}{\epsilon})$ and $\mathcal{O}(q^{\frac{3}{2}}({\log}\;p)^3\sqrt{n}{\log}\;\frac{n}{\epsilon})$ iteration complexity for large- and small-update methods, respectively. These are currently the best known complexity results for such methods.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.60 no.12
• /
• pp.2352-2360
• /
• 2011

In this paper, we introduce a comprehensive design methodology of Radial Basis Function Neural Networks (RBFNN) that is based on mechanism of clustering and optimization algorithm. We can divide some clusters based on similarity of input dataset by using clustering algorithm. As a result, the number of clusters is equal to the number of nodes in the hidden layer. Moreover, the centers of each cluster are used into the centers of each receptive field in the hidden layer. In this study, we have applied Fuzzy-C Means(FCM) and K-Means(KM) clustering algorithm, respectively and compared between them. The weight connections of model are expanded into the type of polynomial functions such as linear and quadratic. In this reason, the output of model consists of relation between input and output. In order to get the optimal structure and better performance, Particle Swarm Optimization(PSO) is used. We can obtain optimized parameters such as both the number of clusters and the polynomial order of weights connection through structural optimization as well as the widths of receptive fields through parametric optimization. To evaluate the performance of proposed model, NXT equipment offered by National Instrument(NI) is exploited. The situation awareness system-related intelligent model was built up by the experimental dataset of distance information measured between object and diverse sensor such as sound sensor, light sensor, and ultrasonic sensor of NXT equipment.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.39 no.4
• /
• pp.314-320
• /
• 2011

In this work, optimization of two-stage shock absorption system for lunar lander has been carried out. Because of complexity of impact phenomena of shock absorption system, a 1-D constitutive model is proposed to describe the behavior of shock absorption system. Quadratic polynomial regression meta-model is constructed by using a commercial software ABAQUS with the proposed 1-D constitutive model, and sequential approximate optimization of two-stage shock absorption system has been carried out along with the constructed meta-model. Through the optimization, it is verified that landing impact force on lunar lander can be considerably reduced by changing the cell size and foil thickness of honeycomb structure in two-stage shock absorption system.

• Nuclear Engineering and Technology
• /
• v.55 no.4
• /
• pp.1382-1399
• /
• 2023

Pressure relief valve (PRV) is one of the important control valves used in nuclear power plants, and its sealing performance is crucial to ensure the safety and function of the entire pressure system. For the sealing performance improving purpose, an explicit function that accounts for all design parameters and can accurately describe the relationship between the multi-design parameters and the seal performance is essential, which is also the challenge of the valve seal design and/or optimization work. On this basis, a surrogate model-based design optimization is carried out in this paper. To obtain the basic data required by the surrogate model, both the Finite Element Model (FEM) and the Computational Fluid Dynamics (CFD) based numerical models were successively established, and thereby both the contact stresses of valve static sealing and dynamic impact (between valve disk and nozzle) could be predicted. With these basic data, the polynomial chaos expansion (PCE) surrogate model which can not only be used for inputs-outputs relationship construction, but also produce the sensitivity of different design parameters were developed. Based on the PCE surrogate model, a new design scheme was obtained after optimization, in which the valve sealing stress is increased by 24.42% while keeping the maximum impact stress lower than 90% of the material allowable stress. The result confirms the ability and feasibility of the method proposed in this paper, and should also be suitable for performance design optimizations of control valves with similar structures.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.28 no.4
• /
• pp.191-198
• /
• 2003

The generalized knapsack problem or gknap is the combinatorial optimization problem of optimizing a nonnegative linear function over the integral hull of the intersection of a polynomially separable 0-1 polytope and a knapsack constraint. The knapsack, the restricted shortest path, and the constrained spanning tree problem are a partial list of gknap. More interesting1y, all the problem that are known to have a fully polynomial approximation scheme, or FPTAS are gknap. We establish some necessary and sufficient conditions for a gknap to admit an FPTAS. To do so, we recapture the standard scaling and approximate binary search techniques in the framework of gknap. This also enables us to find a weaker sufficient condition than the strong NP-hardness that a gknap does not have an FPTAS. Finally, we apply the conditions to explore the fully polynomial approximability of the constrained spanning problem whose fully polynomial approximability is still open.

• Journal of Korea Multimedia Society
• /
• v.20 no.8
• /
• pp.1406-1420
• /
• 2017

Artificial neural networks (ANNs) play an important role in the fields of function approximation, prediction, and classification. ANN performance is critically dependent on the input parameters, including the number of neurons in each layer, and the optimal values of weights and biases assigned to each neuron. In this study, we apply the particle swarm optimization method, a popular optimization algorithm for determining the optimal values of weights and biases for every neuron in different layers of the ANN. Several regression models, including general linear regression, Fourier regression, smoothing spline, and polynomial regression, are conducted to evaluate the proposed method's prediction power compared to multiple linear regression (MLR) methods. In addition, residual analysis is conducted to evaluate the optimized ANN accuracy for both training and test datasets. The experimental results demonstrate that the proposed method can effectively determine optimal values for neuron weights and biases, and high accuracy results are obtained for prediction applications. Evaluations of the proposed method reveal that it can be used for prediction and estimation purposes, with a high accuracy ratio, and the designed model provides a reliable technique for optimization. The simulation results show that the optimized ANN exhibits superior performance to MLR for prediction purposes.