• Smart Media Journal
• /
• v.9 no.3
• /
• pp.31-40
• /
• 2020

Currently, black-box-based machine learning algorithms are used to analyze big data in manufacturing. This algorithm has the advantage of having high analytical consistency, but has the disadvantage that it is difficult to interpret the analysis results. However, in the manufacturing industry, it is important to verify the basis of the results and the validity of deriving the analysis algorithms through analysis based on the manufacturing process principle. To overcome the limitation of explanatory power as a result of this machine learning algorithm, we propose a manufacturing big data analysis method using genetic programming. This algorithm is one of well-known evolutionary algorithms, which repeats evolutionary operators such as selection, crossover, mutation that mimic biological evolution to find the optimal solution. Then, the solution is expressed as a relationship between variables using mathematical symbols, and the solution with the highest explanatory power is finally selected. Through this, input and output variable relations are derived to formulate the results, so it is possible to interpret the intuitive manufacturing mechanism, and it is also possible to derive manufacturing principles that cannot be interpreted based on the relationship between variables represented by formulas. The proposed technique showed equal or superior performance as a result of comparing and analyzing performance with a typical machine learning algorithm. In the future, the possibility of using various manufacturing fields was verified through the technique.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2021.06a
• /
• pp.206-206
• /
• 2021

The necessity for appropriate management of water resources infrastructures such as reservoirs, levees, and dikes is increasing due to unexpected hydro-climate irregularities and rising water demands. To meet this need, past studies have focused on advancing theoretical optimization algorithms such as nonlinear programming, dynamic programming (DP), and genetic programming. Yet, the optimally derived theoretical solutions are limited to be directly implemented in making release decisions in the real-world systems for a variety of reasons. This study first aims to comparatively analyze the two prominent optimization methods, DP and evolutionary multi-objective direct policy search (EMODPS), under historical inflow series using K-fold cross validation. A total of six optimization models are formed each with a specific formulation. Then, one of the optimization models was coupled with the actual zone-based hedging rule that has been adopted in practice. The proposed methodology was applied to Boryeong Dam located in South Korea with conflicting objectives between supply and demand. As a result, the EMODPS models demonstrated a better performance than the DP models in terms of proximity to the ideal. Moreover, the incorporation of the real-world policy with the optimal solutions improved in all indices in terms of the supply side, while widening the range of the trade-off between frequency and magnitude measured in the sides of demand. The results from this study once again highlight the necessity of closing the gap between the theoretical solutions with the real-world implementable policies.

• Geomechanics and Engineering
• /
• v.30 no.6
• /
• pp.551-564
• /
• 2022

Fragmenting the rock mass is considered as the most important work in open-pit mines. Ground vibration is the most hazardous issue of blasting which can cause critical damage to the surrounding structures. This paper focuses on developing an explicit model to predict the ground vibration through an multi objective evolutionary polynomial regression (MOGA-EPR). To this end, a database including 79 sets of data related to a quarry site in Malaysia were used. In addition, a gene expression programming (GEP) model and several empirical equations were employed to predict ground vibration, and their performances were then compared with the MOGA-EPR model using the mean absolute error (MAE), root mean square error (RMSE), mean (𝜇), standard deviation of the mean (𝜎), coefficient of determination (R²) and a20-index. Comparing the results, it was found that the MOGA-EPR model predicted the ground vibration more precisely than the GEP model and the empirical equations, where the MOGA-EPR scored lower MAE and RMSE, 𝜇 and 𝜎 closer to the optimum value, and higher R² and a20-index. Accordingly, the proposed MOGA-EPR model can be introduced as a useful method to predict ground vibration and has the capacity to be generalized to predict other blasting effects.

• Journal of Electrical Engineering and Technology
• /
• v.7 no.3
• /
• pp.376-383
• /
• 2012

A number of scientific researches are currently being conducted on the potential health hazards of power frequency electric and magnetic field (EMF). There exists a non-objective and psychological belief that they are harmful, although no scientific and objective proof of such exists. This possible health risk from ELF magnetic field (MF) exposure, especially for children under 17 years of age, is currently one of Korea's most highly contested social issues. Therefore, to assess the magnetic field exposure levels of those children in their general living environments, the personal MF exposure levels of 436 subjects were measured for about 6 years using government funding. Using the measured database, estimation formulas were developed to predict personal MF exposure levels. These formulas can serve as valuable tools in estimating 24-hour personal MF exposure levels without directly measuring the exposure. Three types of estimation formulas were developed by applying evolutionary computation methods such as genetic algorithm (GA) and genetic programming (GP). After tuning the database, the final three formulas with the smallest estimation error were selected, where the target estimation error was approximately 0.03 ${\mu}T$. The seven parameters of each of these three formulas are gender (G), age (A), house type (H), house size (HS), distance between the subject's residence and a power line (RD), power line voltage class (KV), and the usage conditions of electric appliances (RULE).

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2004.04a
• /
• pp.277-280
• /
• 2004

In the research of biomechanical engineering, robotics and neurophysiology, to clarify the mechanism of human bipedal walking is of major interest. It serves as a basis of developing several applications such as rehabilitation tools and humanoid robots Nevertheless, because of complexity of the neuronal system that Interacts with the body dynamics system to make walking movements, much is left unknown about the details of locomotion mechanism. Researchers were looking for the optimal model of the neuronal system by trials and errors. In this paper, we applied Genetic Programming to induce the model of the nervous system automatically and showed its effectiveness by simulating a human bipedal walking with the obtained model.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2005.04a
• /
• pp.189-192
• /
• 2005

공학시스템의 설계 문제는 시스템 구성이 복잡하고, 구조적으로 열려있으며, 전기, 기계, 유압, 열등의 서로 다른 에너지 도메인 구성 요소를 포함한다. 최적의 설계를 위해서는 각 도메인에 대한 통합된 설계 방법과 자동적이고 구조적으로 열린 공간에 대한 효율적인 탐색방법이 요구된다. 본 논문은 도메인에 독립적이며 모델링과 해석에 장점을 가진 본드 그래프 (bond graph)와 대규모 공간 해의 탐색에 접합한 진화 알고리즘의 일종인 유전자 프로그래밍(Genetic Programming)을 결합하여 멀티 도메인 동적시스템에 대한 디자인 해를 자동적으로 생성해주는 설계 방법을 제시하고, 제안된 설계방법의 효용성을 입증하기 위해서 이를 아나로그 필터 설계에 문제에 적용하였다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.10b
• /
• pp.823-825
• /
• 2003

시간이 흐름에 따라 생화학 시스템이 변화하는 것을 기록한 데이터로부터 이 시스템의 상태 전이 및 시스템을 구성하는 각 생화학 물질간의 관계를 모델링하기 위한 방법으로 S-tree 구조를 제안한다. 이것은 주로 생화학 시스템의 동적 특성을 모델링 하기 위하여 연구되어 온 S-system을 나무 구조로 표현한 것이다. 본 논문에서는 진화 연산을 통해 주어진 시계열 데이터를 잘 설명하는 S-tree의 구조 및 그 변수들을 동시에 효과적으로 탐색하는 방법을 개발하였다. 이 방법에서는 구조 탐색을 위해 유전 프로그래밍(genetic programming)에서 사용되어 온 나무 구조의 교차 및 돌연변이 연산과 더불어 다양한 형태의 구조 탐색 연산자들을 도입하였고, 또한 동시에 알맞은 변수 값들을 찾기 위하여 확률적 돌연변이 연산을 통한 언덕 오르기(hill-climbing)를 수행한다. 제안된 방법을 효모의 혐기성 발효 데이터에 적용한 결과 주어진 시스템을 성공적으로 모델링할 수 있었다.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.25 no.5
• /
• pp.477-482
• /
• 2015

Linear regressions and evolutionary nonlinear regression based compensation techniques for the short-range prediction of wind speed are investigated. Development of an efficient MOS(Model Output Statistics) is necessary to correct systematic errors of the model, but a linear regression based MOS is hard to manage an irregular nature of weather prediction. In order to solve the problem, a nonlinear and symbolic regression method using GP(Genetic Programming) is suggested for a development of MOS for wind speed prediction. The proposed method is compared to various linear regression methods for prediction of wind speed. Also, statistical analysis of distribution for UM elements for each method is executed. experiments are performed for KLAPS(Korea Local Analysis and Prediction System) re-analysis data from 2007 to 2013 year for Jeju Island and Busan area in South Korea.

• KIPS Transactions on Software and Data Engineering
• /
• v.2 no.5
• /
• pp.347-352
• /
• 2013

In impulse noise filtering techniques window size play an important role. Usually, an appropriate window is determined according to the noise density. A small window may not be able to suppress noise properly whereas a large window may remove edges and fine image details. Moreover, the value of the central pixel is estimated by considering all pixels within the window. In this work, contrary to the previous approaches, we propose an iterative impulse noise removal scheme that emphasizes on noise-free pixels within a small neighborhood. The iterative process continues until all noisy pixels are replaced with the estimated pixels. In order to estimate the optimal value for a noisy pixel, a genetic programming (GP) based estimator is evolved that takes few noise-free pixels as input. The estimator is constituent of noise-free pixels, arithmetic operators and random constants. Experimental results show that theproposed scheme is capable of removing impulse noise effectively while preserving the fine image details. Especially, our approach has shown effectiveness against high impulse noise density.

• Proceedings of the Korean Society for Bioinformatics Conference
• /
• 2003.10a
• /
• pp.288-297
• /
• 2003

본 논문은 진화 알고리즘(Evolutionary algorithm)의 기법중의 하나인 유전자 프로그래밍(Genetic programming)을 이용하여 miRNA 유전자를 발굴하기 위한 알고리즘을 소개하고 있다 miRNA는 세포내에서 유전자의 전사를 중지시킴으로써 유전자의 발현을 직접적으로 조절하게 되는 작은 RNA 집단 중의 하나이다. 그러므로 miRNA를 유전체 데이터에서 동정해내는 작업은 생물학적으로 상당히 중요하다. 한편 유전체 데이터에서 miRNA를 동정해내는 알고리즘은 생물학적 실험에서의 시간과 비용을 상당히 절감할 수 있으며, 생물학적으로 miRNA를 동정하는 많은 어려움을 덜어주게 된다. 하지만 계산학적으로 miRNA의 동정은 1차 염기서열상의 통계적인 중요도가 부족하여 기존의 유전자 예측 알고리즘을 적용하기에는 어려움이 있다. 따라서 본 연구에서는 miRNA의 염기서열보다는 2차구조에서 더 많은 유사성을 갖는다는 점을 착안하여, 2차구조내에서 공통적인 구조를 찾아내고, 그 정보를 이용하여 miRNA를 동정해내는 방법으로 접근하였다. 이 알고리즘의 성능평가를 위해 우리는 test set을 이용하여 학습된 모델의 특이도(= 34/38)와 민감도(= 38/67)를 계산하였다. 평가결과 본 알고리즘이 기존의 miRNA 예측 프로그램보다 높은 특이도를 갖고 있으며, 유사한 수준의 민감도를 갖고 있음을 보여 주고 있다.