• 한국경영정보학회:학술대회논문집
• /
• 한국경영정보학회 2007년도 International Conference
• /
• pp.608-613
• /
• 2007

Prediction of corporate bankruptcies has long been an important topic and has been studied extensively in the finance and management literature because it is an essential basis for the risk management of financial institutions. Recently, support vector machines (SVMs) are becoming popular as a tool for bankruptcy prediction because they use a risk function consisting of the empirical error and a regularized term which is derived from the structural risk minimization principle. In addition, they don't require huge training samples and have little possibility of overfitting. However. in order to Use SVM, a user should determine several factors such as the parameters ofa kernel function, appropriate feature subset, and proper instance subset by heuristics, which hinders accurate prediction results when using SVM In this study, we propose a novel hybrid SVM classifier with simultaneous optimization of feature subsets, instance subsets, and kernel parameters. This study introduces genetic algorithms (GAs) to optimize the feature selection, instance selection, and kernel parameters simultaneously. Our study applies the proposed model to the real-world case for bankruptcy prediction. Experimental results show that the prediction accuracy of conventional SVM may be improved significantly by using our model.

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2005년도 춘계학술대회논문집
• /
• pp.227-230
• /
• 2005

The paper presents the identification of dynamic property of a rotor system with a squeeze film damper (SFD) using magnetic fluid. An electromagnet is installed in the inner damper of the SFD. The magnetic fluid is well known as a functional fluid. Its rheological property can be changed by controlling the applied current to the fluid and the fluid can be used as lubricant. Basically, the proposed SFD has the characteristics of a conventional SFD without an applied current, while the damping and stiffness properties change according to the variation of the applied electric current. Therefore, when the applied current is changed, the whirling vibration of the rotor system can be effectively reduced. The clustering-based hybrid evolutionary algorithm (CHEA) is used to identify linear stiffness and damping coefficients of the SFD based on measured unbalance responses.

• 한국경영과학회:학술대회논문집
• /
• 한국경영과학회 2001년도 추계학술대회 논문집
• /
• pp.253-256
• /
• 2001

Today's rapid development in the computer and network technology makes the environment which enables the companies to consider their decisions on the wide point of view and enables the software vendors to make the software packages to help these decisions. To make these software packages, many algorithms should be developed. The production and distribution planning problem belongs to those problems that industry manufacturers daily face in organizing their overall production plan. However, this combinatorial optimization problem can not be solved optimally in a reasonable time when large instances are considered. This legitimates the search for heuristic techniques. As one of these heuristic techniques, genetic algorithm has been considered in many researches. A standard genetic algorithm is a problem solving method that apply the rules of reproduction, gene crossover, and mutation to these pseudo-organisms so those organisms can Pass beneficial and survival-enhancing traits to new generation. This standard genetic algorithm should not be applied to this problem directly because when we represent the chromosome of this problem, there may exist high epitasis between genes. So in this paper, we proposed the hybrid genetic algorithm which turns out to better result than standard genetic algorithms

• 한국정밀공학회지
• /
• 제31권10호
• /
• pp.865-871
• /
• 2014

Paradigm shift to 3-D chip stacking in electronic packaging has induced a lot of integration challenges due to the reduction in wafer thickness and pitch size. This study presents a hybrid bonding technology by self-alignment effect in order to improve the flip chip bonding accuracy with ultra-thin wafer. Optimization of Cu pillar bump formation and evaluation of various factors on self-alignment effect was performed. As a result, highly-improved bonding accuracy of thin wafer with a $50{\mu}m$ of thickness was achieved without solder bridging or bump misalignment by applying reflow process after thermo-compression bonding process. Reflow process caused the inherently-misaligned micro-bump to be aligned due to the interface tension between Si die and solder bump. Control of solder bump volume with respect to the chip dimension was the critical factor for self-alignment effect. This study indicated that bump design for 3D packaging could be tuned for the improvement of micro-bonding quality.

• Journal of Electrical Engineering and Technology
• /
• 제11권1호
• /
• pp.47-54
• /
• 2016

Microgrids offer several reliability benefits, such as the improvement of load-point reliability and the opportunity for reliability-differentiated services. The primary goal of this work is to investigate the impacts of operating condition on the reliability index for microgrid system. It relies on a component failure rate model which quantifies the relationship between component failure rate and state variables. Some parameters involved are characterized by subjective uncertainty. Thus, fuzzy numbers are introduced to represent such parameters, and an optimization model based on Fuzzy Chance Constrained Programming (FCCP) is established for reliability index calculation. In addition, we present a hybrid algorithm which combines scenario enumeration and fuzzy simulation as a solution tool. The simulations in a microgrid test system show that reliability indices without considering operating condition can often prove to be optimistic. We also investigate two groups of situations, which include the different penetration levels of microsource and different confidence levels. The results support the necessity of considering operating condition for achieving accurate reliability evaluation.

• 대한산업공학회지
• /
• 제39권5호
• /
• pp.361-366
• /
• 2013

Prediction of process parameters is very important in parameter design. If predictions are fairly accurate, the quality improvement process will be useful to save time and reduce cost. The concept of dual response approach based on response surface methodology has widely been investigated. Dual response approach may take advantages of optimization modeling for finding optimum setting of input factor by separately modeling mean and variance responses. This study proposes an alternative dual response approach based on machine learning techniques instead of statistical analysis tools. A hybrid neural network-genetic algorithm has been proposed for the purpose of parameter design. A neural network is first constructed to model the relationship between responses and input factors. Mean and variance responses correspond to output nodes while input factors are used for input nodes. Using empirical process data, process parameters can be predicted without performing real experimentations. A genetic algorithm is then applied to find the optimum settings of input factors, where the neural network is used to evaluate the mean and variance response. A drug formulation example from pharmaceutical industry has been studied to demonstrate the procedures and applicability of the proposed approach.

• 한국경영과학회지
• /
• 제22권3호
• /
• pp.163-173
• /
• 1997

Although numerous studies demonstrate that one technique outperforms the others for a given data set, it is hard to tell a priori which of these techniques will be the most effective to solve a specific problem. It has been suggested that the better approach to classification problem might be to integrate several different forecasting techniques by combining their results. The issues of interest are how to integrate different modeling techniques to increase the predictive performance. This paper proposes the post-model integration method, which tries to find the best combination of the results provided by individual techniques. To get the optimal or near optimal combination of different prediction techniques, Genetic Algorithms (GAs) are applied, which are particularly suitable for multi-parameter optimization problems with an object function subject to numerous hard and soft constraints. This study applies three individual classification techniques (Discriminant analysis, Logit model and Neural Networks) as base models for the corporate failure prediction. The results of composite predictions are compared with the individual models. Preliminary results suggests that the use of integrated methods improve the performance of business classification.

• Journal of Welding and Joining
• /
• 제15권1호
• /
• pp.125-134
• /
• 1997

This study was conducted by L9 orthogonal array to obtain optimum spray parameters for This study was conducted by L9 orthogonal array to obtain optimum spray parameters for $Cr_3C_2 - 7wt%$(80wt%Ni-20wt%Cr) coating powder. The factors were hydrogen flow rate, oxygen flow rate, gun-to-work distance, powder feed rate. And evaluation methods for the coating were surface roughness, oxygen concentration, micro-hardness, pore size and distribution, low angle ($30^{\circ}$) erosion rate, and microstructure of coating. The optimum HVOF spray conditions were proved as follows : hydroen flow rate ; 681 SLPM, oxygen flow rate ; 215 SLPM $H^2/O^2 ratio= 3.16), gun-to-work distance ; 22cm, powder feed rate; 25g/min. The hardness (Hv300) was 1147 and the erosion rate ($30^{\circ}$degree) was $3.16\times10^{-4}$g/g. It is believed that the optimized spray conditions can be improved the wear-resistance and anti-erosion characteristics of the coating.

• 상하수도학회지
• /
• 제20권3호
• /
• pp.403-410
• /
• 2006

The objective of this study was to evaluate the possibility of sedimentation basin omission when installed hybrid membrane filtration process in the field plant with the capacity of $500m^3/day$ for 11 months in the "G", water purification plant in Seoul. In order to evaluate the possibility of the sedimentation basin omission, we measured the change of DOC by coagulant dosage. Dosage of PAC(power activated carbon) 4mg/L and coagulant($AI_2O_3$ 10%) 1.67mg/L were compatible to meet the water quality. Also according to the experiment without settlement process, optimization G values were determined to be 300/s, 64/s, and 32/s at the mixing tank, the first flocculator and the second flocculator, respectively. The test was performed under the conditions PAC-coagulation-no settlement-MF. As a result, a dosage of 4.0mg/L as PAC and 0.86 to 1.22mg/L as $Al_2O_3$(10%) in the condition of flux of 62.5LMH were determined to keep TMP value less than $1.0kg_f/cm^2$.

• Journal of the Optical Society of Korea
• /
• 제12권1호
• /
• pp.25-30
• /
• 2008

The optical performances of 15-inch edge-lit backlight were simulated by using a Monte Carlo ray-tracing technique. The backlight model was built by combining a wedge-type light guide plate, a diffuser sheet, a tubular fluorescent lamp with a lamp reflector, and two crossed prism sheets. Angular distributions of the luminance on each optical component obtained from simulation were consistent with those obtained from experiments on a real 15-inch backlight. The constructed backlight model was used to evaluate the optical performances of a micro-pyramid film. It was found that the on-axis luminance gain on the pyramid film is higher than that on one prism film but much lower than that on the two crossed prism films. These results suggest that a reliable simulation model can be used to develop new hybrid films and to optimize the optical structure of edge-lit backlight in order to reduce the developmental period.