• 한국군사과학기술학회지
• /
• 제9권1호
• /
• pp.24-32
• /
• 2006

This paper is to analyze the effectiveness of multistatic sonar network based on detection performance. The multistatic sonar network is a distributed detection system that places a source and multi-receivers apart. So it needs a detection technique that relates to decision rule and optimization of sonar system to improve the detection performance. For this we propose a data fusion procedure using Bayesian decision and optimal sensor arrangement by optimizing a bistatic sonar. Also, to analyze the detection performance effectively, we propose the environmental model that simulates a propagation loss and target strength suitable for multistatic sonar networks in real surroundings. The effectiveness analysis on the multistatic sonar network confirms itself as a promising tool for effective allocation of detection resources in multistatic sonar system.

• Fisheries and Aquatic Sciences
• /
• 제24권4호
• /
• pp.139-152
• /
• 2021

The main purpose of this study is to fit catch-per-unit-effort (CPUE) data about Korea chub mackerel (Scomber japonicus) stock with a state-space production (SSP) model, and to provide stock assessment results. We chose a surplus production model for the chub mackerel data, namely annual yield and CPUE. Then we employed a state-space layer for a production model to consider two sources of variability arising from unmodelled factors (process error) and noise in the data (observation error). We implemented the model via script software ADMB-RE because it reduces the computational cost of high-dimensional integration and provides Markov Chain Monte Carlo sampling, which is required for Bayesian approaches. To stabilize the numerical optimization, we considered prior distributions for model parameters. Applying the SSP model to data collected from commercial fisheries from 1999 to 2017, we estimated model parameters and management references, as well as uncertainties for the estimates. We also applied various production models and showed parameter estimates and goodness of fit statistics to compare the model performance. This study presents two significant findings. First, we concluded that the stock has been overexploited in terms of harvest rate from 1999 to 2017. Second, we suggest a SSP model for the smallest goodness of fit statistics among several production models, especially for fitting CPUE data with fluctuations.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제17권4호
• /
• pp.1200-1215
• /
• 2023

With the development of IoT and artificial intelligence, location-based services are getting more and more attention. For solving the current problem that indoor positioning error is large and generalization is poor, this paper proposes a Model Stacking Algorithm for Indoor Positioning System using WiFi fingerprinting. Firstly, we adopt a model stacking method based on Bayesian optimization to predict the location of indoor targets to improve indoor localization accuracy and model generalization. Secondly, Taking the predicted position based on model stacking as the observation value of particle filter, collaborative particle filter localization based on model stacking algorithm is realized. The experimental results show that the algorithm can control the position error within 2m, which is superior to KNN, GBDT, Xgboost, LightGBM, RF. The location accuracy of the fusion particle filter algorithm is improved by 31%, and the predicted trajectory is close to the real trajectory. The algorithm can also adapt to the application scenarios with fewer wireless access points.

• Structural Engineering and Mechanics
• /
• 제37권4호
• /
• pp.427-442
• /
• 2011

In fatigue life design of mechanical components, uncertainties arising from materials and manufacturing processes should be taken into account for ensuring reliability. A common practice is to apply a safety factor in conjunction with a physics model for evaluating the lifecycle, which most likely relies on the designer's experience. Due to conservative design, predictions are often in disagreement with field observations, which makes it difficult to schedule maintenance. In this paper, the Bayesian technique, which incorporates the field failure data into prior knowledge, is used to obtain a more dependable prediction of fatigue life. The effects of prior knowledge, noise in data, and bias in measurements on the distribution of fatigue life are discussed in detail. By assuming a distribution type of fatigue life, its parameters are identified first, followed by estimating the distribution of fatigue life, which represents the degree of belief of the fatigue life conditional to the observed data. As more data are provided, the values will be updated to reduce the credible interval. The results can be used in various needs such as a risk analysis, reliability based design optimization, maintenance scheduling, or validation of reliability analysis codes. In order to obtain the posterior distribution, the Markov Chain Monte Carlo technique is employed, which is a modern statistical computational method which effectively draws the samples of the given distribution. Field data of turbine components are exploited to illustrate our approach, which counts as a regular inspection of the number of failed blades in a turbine disk.

• 공학논문집
• /
• 제3권1호
• /
• pp.65-73
• /
• 1998

영상복원법 중에는 복원하고자 하는 원 영상의 화소밝기분포가 부분적으로 평탄하다는 가정을 한 부가적인 Gibbs 사전정보를 포함하는 방법이 있다. 이 경우 Gibbs 사전정보를 표현하기 위해 원 영상의 화소밝기를 나타내는 실변수 뿐 아니라 경계를 정의하는 이진변수를 포함하는 에너지 함수를 정의하게 된다. 그러나, 이러한 실변수와 이진변수의 복합형태가 존재할 경우 이들을 동시에 추정하는 것은 매우 어려운 것으로 알려져 있다. 본 연구에서는 deterministic annealing 방법의 응용을 고찰하기로 한다. Deterministic annealing 방법은 다른 방법과 달리 실수 값을 취하는 변수 및 이진변수가 복합형태로 존재하는 문제를 다루는데 있어서 매우 원리적이고 효율적인 방법을 제공한다. 이 방법에서는 복합형태를 취하는 원 함수에 근접하도록 하는 일련의 함수들을 정의하게 되는데, 이때 새로운 일련의 함수들은 실변수만을 취하도록 변환된다. 일련의 함수 중 개개의 함수는 조종파라미터(냉각시 온도에 해당)에 의해 지정된다. 고온에서의 에너지 함수는 저온에서의 에너지와 유사하나 좀더 완만한 형태를 취하게 된다. 따라서, 온도를 서서히 낮추면서 고온에서의 에너지 함수를 저온에서의 에너지 함수로 변환시켜 감으로써 에너지 함수를 최소화하는 작업이 용이해 진다. 이것이 연속방법의 핵심이다. 본 논문에서는 이러한 연속방법을 Bayesian 영상복원 모델에 적용하여 그 성능을 실험을 통해 검증한다.

• 한국지반신소재학회논문집
• /
• 제21권3호
• /
• pp.1-10
• /
• 2022

본 논문에서는 지반 구조물의 재료 물성치에 대한 추계장 세트를 이용하여 재료 물성치의 공간적 임의성 및 확률 특성을 역추정한다. 이러한 추정된 재료 물성치의 확률분포 및 확률특성을 이용하여 구조 형상에 대한 위상 최적화를 수행하고, 기존의 결정론적 위상 최적화 결과와 비교한다. 재료 물성치에 대한 한 세트의 추계장들을 생성하고, 각 추계장에서 재료 물성치의 공간적 임의성을 모사한다. 각 추계장에서 재료 물성치의 부분값들을 이용하여 실제 재료 물성치의 확률분포와 확률 특성을 추정한다. 추정된 실제 재료 물성치의 확률특성을 추계장 세트의 확률 특성과 비교한다. 또한, 임의성을 가진 재료탄성계수를 가지는 지반구조물의 최적화 응답변화도와 임의성이 없는 재료탄성계수를 가지는 지반구조의 응답변화도를 비교한다. 따라서, 실제 재료 물성치의 공간적 임의성을 고려한 정량화된 확률론적 위상 최적화 결과를 얻을 수 있다.

• Wind and Structures
• /
• 제36권6호
• /
• pp.423-434
• /
• 2023

Aerodynamic force coefficients are generally obtained from traditional wind tunnel tests or computational fluid dynamics (CFD). Unfortunately, the techniques mentioned above can sometimes be cumbersome because of the cost involved, such as the computational cost and the use of heavy equipment, to name only two examples. This study proposed to build a deep neural network model to predict the aerodynamic force coefficients based on data collected from CFD simulations to overcome these drawbacks. Therefore, a series of CFD simulations were conducted using different geometric parameters to obtain the aerodynamic force coefficients, validated with wind tunnel tests. The results obtained from CFD simulations were used to create a dataset to train a multilayer perceptron artificial neural network (ANN) model. The models were obtained using three optimization algorithms: scaled conjugate gradient (SCG), Bayesian regularization (BR), and Levenberg-Marquardt algorithms (LM). Furthermore, the performance of each neural network was verified using two performance metrics, including the mean square error and the R-squared coefficient of determination. Finally, the ANN model proved to be highly accurate in predicting the force coefficients of similar bridge sections, thus circumventing the computational burden associated with CFD simulation and the cost of traditional wind tunnel tests.

• 소음진동
• /
• 제6권3호
• /
• pp.305-315
• /
• 1996

Present paper aims at the correlation of modal characteristics of folding fin between test and analysis using an optimization theory. Folding fin is composed of a movable fin, a base fin, and many functional components related to the folding mechanism. Joint parts of folding fin in FEM are initially modeled as rigid elements resulting some difference between test and analysis in modal characteristics. Therefore, some equivalent springs representing joint parts are introduced to improve the FEM model. The springs were set as design variables, while the frequency difference between test and analysis was set as the object function. Bayesian procedure was ujsed for the minimization.

• Journal of Information Processing Systems
• /
• 제3권2호
• /
• pp.73-81
• /
• 2007

In this paper, we introduce a domain-independent classification framework based on both k-nearest neighbor and Naive Bayesian classification algorithms. The architecture of our system is simple and modularized in that each sub-module of the system could be changed or improved efficiently. Moreover, it provides various feature selection mechanisms to be applied to optimize the general-purpose classifiers for a specific domain. As for the enhanced classification performance, our system provides conditional probability boosting (CPB) mechanism which could be used in various domains. In the mood classification domain, our optimized framework using the CPB algorithm showed 1% of improvement in precision and 2% in recall compared with the baseline.

• 한국정보과학회:학술대회논문집
• /
• 한국정보과학회 2006년도 한국컴퓨터종합학술대회 논문집 Vol.33 No.1 (A)
• /
• pp.1-3
• /
• 2006

대사증후군은 당뇨병, 고혈압, 복부 비만, 고지혈증 등의 질병이 한 개인에게 동시에 발현하는 것을 말하며, 최근 경제여건의 향상 및 식생활 습관의 변화와 함께 우리나라에서도 심각한 문제가 되고 있다. 한편 불확실성의 처리를 위해 많이 사용되는 베이지안 네트워크는 사람이 분석 가능한 확률 기반의 모델로 최근 의학분야에서 질병의 진단이나 예측모델을 구성하기 위한 방법으로 유용하게 사용되고 있다. 베이지안 네트워크의 구조를 학습하는 대표적인 알고리즘인 K2 알고리즘은 속성이 입력되는 순서의 영향을 받으며, 따라서 이 또한 하나의 주제로써 연구되어 왔다. 본 논문에서는 유전자 알고리즘을 이용하여 베이지안 네트워크에 입력되는 속성 순서를 최적화하며 이 과정에서 의학지식을 적용해 효율적인 최적화가 가능하도록 하였다. 제안하는 모델을 통해 1993년의 데이터를 가지고 1995년의 상태를 예측하는 분류 실험을 수행한 결과 속성 순서 최적화 후에 이전보다 향상된 예측율을 보였으며 또한 다층 신경망, k-최근접 이웃 등을 이용한 다른 모델보다 더 높은 예측율을 보였다.