• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2002.12a
• /
• pp.162-165
• /
• 2002

Q-learning은 강화학습의 한 방법으로서, 여러 분야에 널리 응용되고 있는 기법이다. 최근에는 Linear Quadratic Regulation(이하 LQR) 문제에 성공적으로 적용된 바 있는데, 특히, 시스템모델의 파라미터에 대한 구체적인 정보가 없는 상태에서 적절한 입력과 출력만을 가지고 학습을 통해 문제를 해결할 수 있어서 상황에 따라서 매우 실용적인 대안이 될 수 있다. Neural Q-learning은 이러한 Q-learning의 Q-value를 MLP(multilayer perceptron) 신경망의 출력으로 대치시킴으로써, 비선형 시스템의 최적제어 문제를 다룰 수 있게 한 방법이다. 그러나, Neural Q방식은 신경망의 구조를 먼저 결정한 후 역전파 알고리즘을 이용하여 학습하는 절차를 취하기 때문에, 시행착오를 통하여 신경망 구조를 결정해야 한다는 점, 역전파 알고리즘의 적용으로 인해 신경망의 연결강도 값들이 지역적 최적해로 수렴한다는 점등의 문제점을 상속받는 한계가 있다. 따라서, 본 논문에서는 Neural-0 학습의 도구로, 역전파 알고리즘으로 학습되는 MLP 신경망을 사용하는 대신 최근 들어 여러 분야에서 그 성능을 인정받고 있는 서포트 벡터 학습법을 사용하는 방법을 택하여, $\varepsilon$-SVR(Epsilon Support Vector Regression)을 이용한 Q-value 근사 기법을 제안하고 관련 수식을 유도하였다. 그리고, 모의 실험을 통하여, 제안된 서포트 벡터학습 기반 Neural-Q 방법의 적용 가능성을 알아보았다.

• Journal of the Korean Society of Radiology
• /
• v.84 no.6
• /
• pp.1211-1219
• /
• 2023

Korea is rapidly entering into an aging society, and an increasing socioeconomic burden related to prostate cancer is inevitable. Therefore, the need for early detection and accurate diagnosis of prostate cancer is becoming increasingly critical. Ideally, a biopsy should accurately detect cancers using a minimum number of cores. However, as prostate cancer is often indistinguishable on imaging, image-guided targeted biopsies alone are insufficient for diagnosis. After decades of trial and error, the diagnosis of prostate cancer relies heavily on systematic biopsy, which is characterized by random and repetitive core acquisition throughout the gland. This review will provide an overview of the historical aspects of prostate cancer diagnosis. Moreover, the review will also address the general considerations involved in prostate biopsy, and discuss the periprocedural management of the patients.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.21 no.6
• /
• pp.411-416
• /
• 2020

An optical arrangement has been set inside a photo-spectrometer to measure the reflectance of IR optics materials in mid IR range. The optical arrangement consists of equally spaced 4 gold coated full reflecting mirrors with the incidence angle of 45°. Baseline beam intensity I_B has been measured while the beam proceeds through the 4 mirrors. Reflectance of a mirror has been estimated from the I_B. And the beam intensity I_S with the specimen in the optical path has been measured with the 4th mirror replaced with the specimen. Reflectance of the specimen has been estimated from the value of I_S/I_B. Then the estimated reflectance has been put in Fresnel equation relating reflectance and refractive index(RI) to estimate the RI of the material. Measurement has been made for sapphire, germanium, magnesium fluoride, and zinc sulfide. The estimated RI of the materials are closely matching with reference data and the maximum difference less than 2% over the wavelength range 3-5㎛ for all materials tested. As an FT-IR photo-spectrometer with a broadband wavelength infrared light source is used, this method has the advantage of measuring the refractive index at multiple wavelengths in a single measurement.

• Journal of the Korea Computer Graphics Society
• /
• v.14 no.4
• /
• pp.35-39
• /
• 2008

A computer supported collaborative system provides with a shared virtual workspace over the Internet where its remote users cooperate in order to achieve their goals by overcoming problems caused by distance and time. VRMMS (Virtual Reality Molecular Modeling System) [1] is a VR based collaborative system where biologists can remotely participate in and exercise molecular modeling tasks such as viewing three dimensional structures of molecular models, confirming results of molecular simulations and providing with feedbacks for the next simulations. Biologists can utilize VRMMS in executing molecular simulations. However, first-time users and beginners need to spend some time for studying and practicing in order to skillfully manipulate molecular models and the system. The best way to resolve the problem is to have a face-to-face session of teaching and learning VRMMS. However, it is not practically recommended in the sense that the users are remotely located. It follows that the learning time could last longer than desired. In this paper, we propose to use Second Life [2] combining with VRMMS for removing the problem. It can be used in building a shared workplace over the Internet where molecular simulations using VRMMS can be exercised, taught, learned and practiced. Through the web, users can collaborate with each other using VRMMS. Their avatars and tools of molecular simulations can be remotely utilized in order to provide with senses of 'being there' to the remote users. The users can discuss, teach and learn over the Internet. The shared workspaces for discussion and education are designed and implemented in Second Life. Since the activities in Second Life and VRMMS are designed to realistic, the system is expected to help users in improving their learning and experimental performances.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.18 no.6
• /
• pp.569-579
• /
• 2016

In order to ensure the stability of tunnel portal slope, reinforcement method such as anchors, soil nails and rock bolts have been used in Korea. When selecting slope reinforcement methods in tunnel portal area such as reinforcement arrangement and length, trial and error method can be very time-consuming and it was also not easy to verify the selection of an optimum condition. In this study, using the FISH language embedded in the finite difference code FLAC3D program, the optimization technique was developed with the Differential Evolution Algorithm (DEA). After building a database on the soil nailing method in tunnel portal area, this system can be selected to an optimum arrangement plan based on the factor of safety through the FLAC3D analysis. Through the results of numerical analysis, it was confirmed that the number of analysis was decreased by about 8 times when DEA based optimization technique was used compared to the full combination (FC). In case of the design of slope reinforcement in tunnel portal area, if this built-system is used, it is expected that the selection of an optimum arrangement plan can be relatively easier.

• Journal of Korean Society for Geospatial Information Science
• /
• v.18 no.1
• /
• pp.11-20
• /
• 2010

The distributed rainfall-runoff model which is developed in the country requires a lot of time and effort to generate input data. Also, it takes a lot of time to calculate discharge by numerical analysis based on kinematic wave theory in runoff process. Therefore, most river basins using the distributed model are of limited scale, such as small river basins. However, recently, the necessity of integrated watershed management has been increasing due to change of watershed management concept and discharge calculation of whole river basin, including upstream and downstream of dam. Thus, in this study, the feasibility of the GIS based physical distributed rainfall-runoff model, K-DRUM(K-water hydrologic & hydraulic Distributed RUnoff Model) which has been developed by own technology was reviewed in the flood discharge process for the Geum River basin, including Yongdam and Daecheong Dam Watersheds. GIS hydrological parameters were extracted from basic GIS data such as DEM, land cover and soil map, and used as input data of the model. Problems in running time and inaccuracy setting using the existing trial and error method were solved by applying an auto calibration method in setting initial soil moisture conditions. The accuracy of discharge analysis for application of the method was evaluated using VER, QER and Total Error in case of the typhoon 'Ewiniar' event. and the calculation results shows a good agreement with observed data.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.12 no.3
• /
• pp.127-137
• /
• 1992

For the real-time control of a multi-purpose reservoir in case of a storm, it is absolutely necessary to forecast accurate flood inflows through a good rainfall-runoff model by calibrating the parameters with the on-line rainfall and water level data transmitted by the telemetering systems. To calibrate the parameters of a runoff model. the trial and error method of manual calibration has been adopted from the subjective view point of a model user. The object of this study is to develop a automatic calibration method using an optimization technique. The pattern-search algorithm was applied as an optimization technique because of the stability of the solution under various conditions. The object function was selected as the sum of the squares of differences between observed and fitted ordinates of the hydrograph. Two historical flood events were applied to verify the developed technique for the automatic calibration of the parameters of the storage-function rainfall-runoff model which has been used for the flood control of the Soyanggang multi-purpose reservoir by the Korea Water Resources Corporation. The developed method was verified to be much more suitable than the manual method in flood forecasting and real-time reservoir controlling because it saves calibration time and efforts in addition to the better flood forecasting capability.

• Journal of Navigation and Port Research
• /
• v.44 no.3
• /
• pp.187-194
• /
• 2020

It is very important to forecast freight volume accurately to establish major port policies and future operation plans. Thus, related studies are being conducted because of this importance. In this paper, stepwise regression analysis and artificial neural network model were analyzed to compare the predictive power of each model on Gwangyang Port, the largest domestic port for coal and iron ore transportation. Data of a total of 121 months J anuary 2009-J anuary 2019 were used. Factors affecting coal and iron ore trade volume were selected and classified into supply-related factors and market/economy-related factors. In the stepwise regression analysis, the tonnage of ships entering the port, coal price, and dollar exchange rate were selected as the final variables in case of the Gwangyang Port coal volume forecasting model. In the iron ore volume forecasting model, the tonnage of ships entering the port and the price of iron ore were selected as the final variables. In the analysis using the artificial neural network model, trial-and-error method that various Hyper-parameters affecting the performance of the model were selected to identify the most optimal model used. The analysis results showed that the artificial neural network model had better predictive performance than the stepwise regression analysis. The model which showed the most excellent performance was the Gwangyang Port Coal Volume Forecasting Artificial Neural Network Model. In comparing forecasted values by various predictive models and actually measured values, the artificial neural network model showed closer values to the actual highest point and the lowest point than the stepwise regression analysis.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.22 no.5
• /
• pp.1-18
• /
• 2023

With the recent developments in big data and deep learning, a variety of traffic information is collected widely and used for traffic operations. In particular, long short-term memory (LSTM) is used in the field of traffic information prediction with time series characteristics. Since trends, seasons, and cycles differ due to the nature of time series data input for an LSTM, a trial-and-error method based on characteristics of the data is essential for prediction models based on time series data in order to find hyperparameters. If a methodology is established to find suitable hyperparameters, it is possible to reduce the time spent in constructing high-accuracy models. Therefore, in this study, a traffic information prediction model is developed based on highway vehicle detection system (VDS) data and LSTM, and an impact assessment is conducted through changes in the LSTM evaluation indicators for each hyperparameter. In addition, a methodology for finding hyperparameters suitable for predicting highway traffic information in the transportation field is presented.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.37 no.11
• /
• pp.1397-1405
• /
• 2013

The optimal rib pattern design of B-pillar trim considering occupant head protection can be determined by two methods. One is the conventional approximate optimization method that uses the statistical design of experiments (DOE) and response surface method (RSM). Generally, approximated optimum results are obtained through the iterative process by trial-and-error. The quality of results strongly depends on the factors and levels assigned by a designer. The other is a methodology derived from previous work by the authors, called the sequential design of experiments (SDOE), to reduce the trial-and-error procedure and to find an appropriate condition for using artificial neural network (ANN) systematically. An appropriate condition is determined from the iterative process based on the analysis of means. With this new technique and ANN, it is possible to find an optimum design accurately and efficiently.