• 상하수도학회지
• /
• 제35권4호
• /
• pp.259-275
• /
• 2021

This study explored the usefulness and implications of the Bayesian hyperparameter optimization in developing species distribution models (SDMs). A variety of machine learning (ML) algorithms, namely, support vector machine (SVM), random forest (RF), boosted regression tree (BRT), XGBoost (XGB), and Multilayer perceptron (MLP) were used for predicting the occurrence of four benthic macroinvertebrate species. The Bayesian optimization method successfully tuned model hyperparameters, with all ML models resulting an area under the curve (AUC) > 0.7. Also, hyperparameter search ranges that generally clustered around the optimal values suggest the efficiency of the Bayesian optimization in finding optimal sets of hyperparameters. Tree based ensemble algorithms (BRT, RF, and XGB) tended to show higher performances than SVM and MLP. Important hyperparameters and optimal values differed by species and ML model, indicating the necessity of hyperparameter tuning for improving individual model performances. The optimization results demonstrate that for all macroinvertebrate species SVM and RF required fewer numbers of trials until obtaining optimal hyperparameter sets, leading to reduced computational cost compared to other ML algorithms. The results of this study suggest that the Bayesian optimization is an efficient method for hyperparameter optimization of machine learning algorithms.

• Nuclear Engineering and Technology
• /
• 제49권2호
• /
• pp.434-441
• /
• 2017

Containment venting is one of several essential measures to protect the integrity of the final barrier of a nuclear reactor during severe accidents, by which the uncontrollable release of fission products can be avoided. The authors seek to develop an optimization approach to venting operations, from a simulation-based perspective, using an integrated severe accident code, THALES2/KICHE. The effectiveness of the containment-venting strategies needs to be verified via numerical simulations based on various settings of the venting conditions. The number of iterations, however, needs to be controlled to avoid cumbersome computational burden of integrated codes. Bayesian optimization is an efficient global optimization approach. By using a Gaussian process regression, a surrogate model of the "black-box" code is constructed. It can be updated simultaneously whenever new simulation results are acquired. With predictions via the surrogate model, upcoming locations of the most probable optimum can be revealed. The sampling procedure is adaptive. Compared with the case of pure random searches, the number of code queries is largely reduced for the optimum finding. One typical severe accident scenario of a boiling water reactor is chosen as an example. The research demonstrates the applicability of the Bayesian optimization approach to the design and establishment of containment-venting strategies during severe accidents.

• 로봇학회논문지
• /
• 제19권1호
• /
• pp.45-52
• /
• 2024

This paper introduces an approach to enhance the performance of magnetic manipulation systems for microrobot actuation. A variety of eight-electromagnet configurations have been proposed to date. The previous study revealed that achieving 5 degrees of freedom (5-DOF) control necessitates at least eight electromagnets without encountering workspace singularities. But so far, the research considering the influence of iron cores embedded in electromagnets has not been conducted. This paper offers a novel approach to optimizing electromagnet configurations that effectively consider the influence of iron cores. The proposed methodology integrates probabilistic optimization with finite element methods (FEM), using Bayesian Optimization (BO). The Bayesian optimization aims to optimize the worst-case magnetic force generation for enhancing the performance of magnetic manipulation system. The proposed simulation-based model achieves approximately 20% improvement compared to previous systems in terms of actuation performance. This study has the potential for enhancing magnetic manipulation systems for microrobot control, particularly in medical and microscale technology applications.

• Smart Structures and Systems
• /
• 제31권1호
• /
• pp.57-68
• /
• 2023

Bridge hangers, such as those in suspension and cable-stayed bridges, suffer from cumulative fatigue damage caused by dynamic loads (e.g., cyclic traffic and wind loads) in their service condition. Thus, the identification of damage to hangers is important in preserving the service life of the bridge structure. This study develops a new method for condition assessment of bridge hangers. The tension force of the bridge and the damages in the element level can be identified using the Bayesian optimization method. To improve the number of observed data, the additional mass method is combined the Bayesian optimization method. Numerical studies are presented to verify the accuracy and efficiency of the proposed method. The influence of different acquisition functions, which include expected improvement (EI), probability-of-improvement (PI), lower confidence bound (LCB), and expected improvement per second (EIPC), on the identification of damage to the bridge hanger is studied. Results show that the errors identified by the EI acquisition function are smaller than those identified by the other acquisition functions. The identification of the damage to the bridge hanger with various types of boundary conditions and different levels of measurement noise are also studied. Results show that both the severity of the damage and the tension force can be identified via the proposed method, thereby verifying the robustness of the proposed method. Compared to the genetic algorithm (GA), particle swarm optimization (PSO), and nonlinear least-square method (NLS), the Bayesian optimization (BO) performs best in identifying the structural damage and tension force.

• 마이크로전자및패키징학회지
• /
• 제31권3호
• /
• pp.50-57
• /
• 2024

최근 대리 모델에 머신 러닝 기술을 접목하여 복잡한 설계에 대한 최적화를 빠르게 달성하는 방법론이 활발히 연구되고 있다. 훈련된 머신 러닝 대리 모델은 복잡한 유한요소해석 시뮬레이션 대비 컴퓨팅 자원을 적게 소모하면서 동일한 해석 결과를 출력할 수 있다. 또한 훈련된 모델에 최적화를 결합하면 반복 시뮬레이션 대비 더 빠르게 최적의 설계 변수를 도출할 수 있다. 본 연구에서는 패키지 휨을 최소화하는 설계 변수 조합을 효과적으로 탐색하기 위하여 심층신경망과 베이지안 최적화를 적용하였다. 심층신경망 모델은 유한요소해석 시뮬레이션으로 획득한 설계 변수-휨 데이터셋을 바탕으로 훈련하였고, 해당 모델에 베이지안 최적화를 적용하여 휨을 최소화하는 최적의 설계 변수를 탐색하였다. 구축한 심층신경망 및 베이지안 최적화 모델은 실제 시뮬레이션 결과와 99% 이상 일치하는 동시에, 최적 설계 변수 탐색에 소요되는 시간은 15초에 불과하여, 1회의 시뮬레이션과 비교해도 57% 이상 최적화 시간을 단축할 수 있다.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2018년도 학술발표회
• /
• pp.176-176
• /
• 2018

This study presents a regional, probabilistic framework for estimating streamflow via spatial scaling in the Great Lakes basin, which is the largest lake system in the world. The framework follows a two-fold strategy including (1) a quadratic-programming based optimization model a priori to explore the model structure, and (2) a time-varying hierarchical Bayesian model based on insights found in the optimization model. The proposed model is developed to explore three innovations in hierarchical modeling for reconstructing historical streamflow at ungaged sites: (1) information of physical characteristics is utilized in spatial scaling, (2) a time-varying approach is introduced based on climate information, and (3) heteroscedasticity in residual errors is considered to improve streamflow predictive distributions. The proposed model is developed and calibrated in a hierarchical Bayesian framework to pool regional information across sites and enhance regionalization skill. The model is validated in a cross-validation framework along with four simpler nested formulations and the optimization model to confirm specific hypotheses embedded in the full model structure. The nested models assume a similar hierarchical Bayesian structure to our proposed model with their own set of simplifications and omissions. Results suggest that each of three innovations improve historical out-of-sample streamflow reconstructions although these improvements vary corrsponding to each innovation. Finally, we conclude with a discussion of possible model improvements considered by additional model structure and covariates.

• 한국컴퓨터정보학회:학술대회논문집
• /
• 한국컴퓨터정보학회 2022년도 제66차 하계학술대회논문집 30권2호
• /
• pp.53-56
• /
• 2022

본 논문에서는 심전도(ECG) 센서와 에지 컴퓨팅(Edge computing)을 활용하여 실시간 데이터와 Bayesian optimization을 통한 기계학습 알고리즘으로 재활 로봇에서 발목을 제어할 수 있는 Parameter(외골격 관련) 최적값을 출력한다. 심전도 센서 적용을 기반으로 하는 바이오 데이터 기술, 기계 학습(Bayesian optimization) 모델 접근 방식과 하드웨어 결합으로 재활 로봇 모터를 제어할 수 있는 Parameter 제공과 실시간 모터 제어 운영할 수 있도록 분석 플랫폼을 구축한다. 이 플랫폼을 이용해보다 효과적인 이동형 로봇설계 및 처리 방법을 연결할 수 있는 발판을 마련하였고, 로봇제어에 많이 사용하고 있는 매트랩 시뮬링크(Matlab simulink)를 연결할 수 있는 범용 통신 지원한다. 센서-전처리-인공지능 알고리즘-모터 제어 Parameter로 연계되는 데이터 가공과 처리 방법으로 최근 분석 기법을 적용하여 바이오 데이터 연구 활동과 이동형 재활 로봇 관련 데이터 분석 분야를 쉽게 접근할 수 있도록 한다.

• 한국수자원학회논문집
• /
• 제45권5호
• /
• pp.505-516
• /
• 2012

본 연구에서는 국내외에서 범용되고 있는 단일강우사상 모형인 미육군공병단의 HEC-1 모형을 이용하여 대청댐 유역의 실측 강우-유출 사상을 중심으로 강우-유출 모의를 수행하였으며, 매개변수 검정에는 실제 대청댐의 시간당 유입량을 기준으로 검정을 실시하였다. HEC-1 모형에는 매개변수를 자동으로 최적화시키는 프로그램이 내장되어 있으나 본 연구의 대상유역과 같이 다수의 소유역이 있는 경우, 매개변수 추정시 매개변수 중 일부는 수렴되지 못하고 발산하는 문제가 있었으며, 첨두유량의 추정능력 역시 저하되는 문제를 보였다. 따라서 이러한 HEC-1 모형의 매개변수의 불확실성을 고려하기 위한 방안으로 Bayesian 모형을 HEC-1모형에 연동시켜 활용하였으며, 기존 HEC-1 강우-유출 모형에 적용할 수 있는 매개변수 최적화 및 불확실성 정량화를 위해 HEC-1 강우-유출 모형 매개변수는 SCS 1개, Clark 단위도 2개를 Bayesian MCMC 기법을 적용하여 매개변수간 조건부확률로 모의발생을 한 후, Bayesian 모형으로부터 각 매개변수의 사후분포(posterior distribution)를 추정하여 사후분포의 추정이 매개변수의 불확실성 정량화를 수행하였다. 본 연구를 통해 제안된 BHEC-1 모형을 대상으로 대청댐 유역에 실측 강우-유출사상에 대해서 모형의 적합성을 평가한 결과, 7개 유역의 21개의 매개변수가 해의 발산 없이 안정된 매개변수 추정이 가능하였다. 한편, Bayesian 모형을 근간으로 하기 때문에 최종결과로서 매개변수들의 사후분포(posterior)의 추정이 가능하여 향후 홍수빈도곡선 유도, 댐 위험도분석과 기후변화 문제와 같은 다양한 수문학적 문제의 연구에 적용 가능할 것으로 전망된다.

• ETRI Journal
• /
• 제33권1호
• /
• pp.39-49
• /
• 2011

A new structure learning approach for Bayesian networks based on asexual reproduction optimization (ARO) is proposed in this paper. ARO can be considered an evolutionary-based algorithm that mathematically models the budding mechanism of asexual reproduction. In ARO, a parent produces a bud through a reproduction operator; thereafter, the parent and its bud compete to survive according to a performance index obtained from the underlying objective function of the optimization problem: This leads to the fitter individual. The convergence measure of ARO is analyzed. The proposed method is applied to real-world and benchmark applications, while its effectiveness is demonstrated through computer simulations. Results of simulations show that ARO outperforms genetic algorithm (GA) because ARO results in a good structure and fast convergence rate in comparison with GA.

• 한국군사과학기술학회지
• /
• 제25권1호
• /
• pp.72-81
• /
• 2022

Maritime patrol aircraft is an efficient solution for detecting submarines at sea. The aircraft can only detect submarines by sonobuoy, but the number of buoy is limited. In this paper, we present the online sonobuoy deployment method for estimating the location of submarines. We use Gaussian process regression to estimate the submarine existence probability map, and Bayesian optimization to decide the next best position of sonobuoy. Further, we show the performance of the proposed method by simulation.