• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.26 no.12
• /
• pp.2556-2564
• /
• 2002

The design of airfoil had practiced by repeat tests in its first stage, though an airfoil has as been designed based on simulations according to techniques of computational fluid dynamics. Here, using of traditional optimization is unsuitable because a state of flux is hypersensitive to the shape of airfoil. Therefore the paper optimized the shape of airfoil in transonic region using a genetic algorithm (GA). Response surfaces are based on back propagation neural network (BPN) and regression model. Training data of BPN and regression model were obtained by computational fluid dynamic analysis using CFD-ACE, and each analysis has been designed by design of experiments.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1997.11a
• /
• pp.267-270
• /
• 1997

In this paper, we proposed an optimazation method using Genetic Algorithm for nonlinear system modeling. Fuzzy Neural Network(FNNs) was used as basic model of nonlinear system. FNNs was fused of Fuzzy Inference which has linguistic property and Neural Network which has learning ability and high tolerence level. This paper, We used FNNs which was proposed by Yamakawa. The FNNs was composed Simple Inference and Error Back Propagation Algorithm. To obtain optimal model, parameter of membership function, learning rate and momentum coefficient of FNNs are tuned using genetic algorithm. And we used simplex algorithm additionaly to overcome limit of genetic algorithm. For the purpose of evaluation of proposed method, we applied proposed method to traffic choice process and waste water treatment process, and then obtained more precise model than other previous optimization methods and objective model.

• Journal of Electrical Engineering and Technology
• /
• v.12 no.3
• /
• pp.989-995
• /
• 2017

As is known to all that the output of wind power generation has a character of randomness and volatility because of the influence of natural environment conditions. At present, the research of wind power prediction mainly focuses on point forecasting, which can hardly describe its uncertainty, leading to the fact that its application in practice is low. In this paper, a wind power range prediction model based on the multiple output property of BP neural network is built, and the optimization criterion considering the information of predicted intervals is proposed. Then, improved Particle Swarm Optimization (PSO) algorithm is used to optimize the model. The simulation results of a practical example show that the proposed wind power range prediction model can effectively forecast the output power interval, and provide power grid dispatcher with decision.

• Proceedings of the KIEE Conference
• /
• 1996.07b
• /
• pp.1344-1346
• /
• 1996

In this paper, we discuss an informal analysis of diffusion, global optimization properties of Langevine competitive learning neural network. In the view of the stochastic process, it is important that competitive learning gurantee an optimal solution for pattern recognition. We show that the binary reinforcement function in Langevine competitive learning is a brownian motion as Gaussian process, and construct the Fokker-Plank equation for the proposed neural network. Finally, we show that the informal analysis of the proposed algorithm has a possiblity of globally optimal. solution with the proper initial condition.

• Journal of the Microelectronics and Packaging Society
• /
• v.31 no.3
• /
• pp.50-57
• /
• 2024

Recently, applying a machine learning to surrogate modeling for rapid optimization of complex designs have been widely researched. Once trained, the machine learning surrogate model can predict similar outputs to Finite Element Analysis (FEA) simulations but require significantly less computing resources. In addition, combined with optimization methodologies, it can identify optimal design variable with less time requirement compared to iterative simulation. This study proposes a Deep Neural Network (DNN) model with Bayesian Optimization (BO) approach for efficiently searching the optimal design variables to minimize the warpage of electronic package. The DNN model was trained by using design variable-warpage dataset from FEA simulation, and the Bayesian optimization was applied to find the optimal design variables which minimizing the warpage. The suggested DNN + BO model shows over 99% consistency compared to actual simulation results, while only require 15 second to identify optimal design variable, which reducing the optimization time by more than 57% compared to FEA simulation.

• 제어로봇시스템학회:학술대회논문집
• /
• 1992.10b
• /
• pp.564-568
• /
• 1992

A neural optimization network is designed to solve the collsion-free inverse kinematics problem for redundant robot manipulators under the constraints of joint limits, maximum velocities and maximum accelerations. And the fuzzy rules are proposed to determine the weightings of neural optimization networks to avoid the collision between robot manipulator and obstacles. The inputs of fuzzy rules are the resultant distance, change of the distance and sum of the changes. And the output of fuzzy rules is defined as the capability of collision avoidance of joint differential motion. The weightings of neural optimization networks are adjusted according to the capability of collision avoidance of each joint. To show the validities of the proposed method computer simulation results are illustrated for the redundant robot with three degrees of freedom,

• Journal of Korea Multimedia Society
• /
• v.20 no.8
• /
• pp.1406-1420
• /
• 2017

Artificial neural networks (ANNs) play an important role in the fields of function approximation, prediction, and classification. ANN performance is critically dependent on the input parameters, including the number of neurons in each layer, and the optimal values of weights and biases assigned to each neuron. In this study, we apply the particle swarm optimization method, a popular optimization algorithm for determining the optimal values of weights and biases for every neuron in different layers of the ANN. Several regression models, including general linear regression, Fourier regression, smoothing spline, and polynomial regression, are conducted to evaluate the proposed method's prediction power compared to multiple linear regression (MLR) methods. In addition, residual analysis is conducted to evaluate the optimized ANN accuracy for both training and test datasets. The experimental results demonstrate that the proposed method can effectively determine optimal values for neuron weights and biases, and high accuracy results are obtained for prediction applications. Evaluations of the proposed method reveal that it can be used for prediction and estimation purposes, with a high accuracy ratio, and the designed model provides a reliable technique for optimization. The simulation results show that the optimized ANN exhibits superior performance to MLR for prediction purposes.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.36 no.2
• /
• pp.99-104
• /
• 2008

This study performed robust design optimization of fighter wing planform, considering uncertainty based on neural network model. To construct uncertainty model, aerodynamic performance and their sensitivity were evaluated by 3-dimensional Euler equations and adjoint variable method at experimental points selected from central composite design. In addition, because a neural network model has the advantage of capturing non-linear characteristic, it was possible to predict sensitivity of the aerodynamic performance efficiently and accurately . From the results of robust design optimization, it could be confirmed that the robustness of the objective function and constraints were improved if the variation of uncertainty and sigma level were increased.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.22 no.1
• /
• pp.40-49
• /
• 2008

Interior permanent magnet synchronous motor(IPMSM) has become a popular choice in electric vehicle applications and so of due to their excellent power to weight ratio. To obtain maximum efficiency in these applications, this paper proposes the neural network control method. The controllable electrical loss which consists of the copper loss and the iron loss can be minimized by the error back propagation algorithm(EBPA) of neural network. The minimization of loss is possible to realize eHciency optimization control for the IPMSM drive. This paper proposes high performance and robust control through a real time calculation of parameter variation such as variation of back emf constant, armature resistance and d-axis inductance about the motor operation. Proposed algorithm is applied IPMSM drive system, prove validity through analysis operating characteristics con011ed by efficiency optimization control.

• Journal of Korea Multimedia Society
• /
• v.24 no.4
• /
• pp.519-527
• /
• 2021

Alzheimer's disease is one of the challenges to tackle in the coming aging era and is attempting to diagnose and predict through various biomarkers. While the application of various deep learning-based technologies as powerful imaging technologies has recently expanded across the medical industry, empirical design is not easy because there are various deep earning neural networks architecture and categorical hyperparameters that rely on problems and data to solve. In this paper, we show the possibility of optimizing a deep learning neural network structure and hyperparameters for Alzheimer's disease classification in amyloid brain images in a representative deep earning neural networks architecture using genetic algorithms. It was observed that the optimal deep learning neural network structure and hyperparameter were chosen as the values of the experiment were converging.