• Journal of the Korean Society of Marine Environment & Safety
• /
• v.23 no.3
• /
• pp.258-265
• /
• 2017

This paper validates a system identification method using mathematical optimization using sea trial measurement data as a benchmark. A fast time simulation tool, SIMOPT, and a Rheinmetall Defence mathematical model have been adopted to conduct initial hydrodynamic coefficient estimation and simulate ship modelling. Calibration for the environmental effect of sea trial measurement and sensitivity analysis have been carried out to enable a simple and efficient optimization process. The optimization process consists of three steps, and each step controls different coefficients according to the corresponding manoeuvre. Optimization result of Step 1, an optimization for coefficient on x-axis, was similar compared to values applying an empirical regression formulae by Clarke and Norrbin, which is used for SIMOPT. Results of Steps 2 and 3, which are for linear coefficients and nonlinear coefficients, respectively, was differ from the calculation results of the method by Clarke and Norrbin. A comparison for ship trajectory of simulation results from the benchmark and optimization results indicated that the suggested stepwise optimization method enables a coefficient tuning in a mathematical way.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.30 no.7 s.250
• /
• pp.603-611
• /
• 2006

This paper presents the response surface optimization method using three-dimensional Wavier-Stokes analysis to optimize the blade shape of an axial flow fan. Reynolds-averaged Wavier-Stokes equations with $k-{\epsilon}$ turbulence model are discretized with finite volume approximations using the unstructured grid. Regression analysis is used for generating response surface, and it is validated by ANOVA and t-statistics. Four geometric variables, i.e., sweep and lean angles at mean and tip respectively were employed to improve the efficiency. The computational results are compared with experimental data and the comparisons show generally good agreements. As a main result of the optimization, the total efficiency was successfully improved. Also, detailed effects of sweep and lean on the axial flow fan are discussed.

• 유체기계공업학회:학술대회논문집
• /
• 2004.12a
• /
• pp.578-582
• /
• 2004

This paper presents the response surface optimization method using three-dimensional Navier-Stokes Analysis to optimize the shape of a axial flow fan. Reynolds-averaged Navier-Stokes equations with k-$\epsilon$ turbulence model are discretized with finite volume approximations. Regression analysis is used for generating response surface, and it is validated by ANOVA. Five geometric variables, i.e., distribution of sweep angle at mean and tip, lean angle at mean and tip, and spanwise location of mean were employed to optimize the efficiency. The computational results are compared with experiment data. As a main result of the optimization, the efficiency was successfully improved.

• Transactions of the Korean Society of Automotive Engineers
• /
• v.11 no.4
• /
• pp.94-101
• /
• 2003

Recently, miniaturization and weight reduction is getting more attention due to various benefits in automotive components design. It is a trend that the design of experiment(DOE) and statical design method are frequently used for optimization. In this research, the safety of lower control arm is evaluated according to its material change form S45C to A16061 for the reduction of arm's weight. The variance analysis on the basis of structure analysis and DOE is applied to the lower control m. We have proposed a statistical design model to evaluate the effect of structural modification by performing the practical multi-objective optimization considering mass, stress and deflection.

• Communications for Statistical Applications and Methods
• /
• v.29 no.6
• /
• pp.629-640
• /
• 2022

Variable selection is one of the most crucial tasks in supervised learning, such as regression and classification. The best subset selection is straightforward and optimal but not practically applicable unless the number of predictors is small. In this article, we propose directly solving the best subset selection via the genetic algorithm (GA), a popular stochastic optimization algorithm based on the principle of Darwinian evolution. To further improve the variable selection performance, we propose to run multiple GA to solve the best subset selection and then synthesize the results, which we call ensemble GA (EGA). The EGA significantly improves variable selection performance. In addition, the proposed method is essentially the best subset selection and hence applicable to a variety of models with different selection criteria. We compare the proposed EGA to existing variable selection methods under various models, including linear regression, Poisson regression, and Cox regression for survival data. Both simulation and real data analysis demonstrate the promising performance of the proposed method.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.18 no.2
• /
• pp.299-304
• /
• 2023

Among artificial intelligence techniques, deep learning is a model that has been used in many places and has proven its effectiveness. However, deep learning models are not used effectively in everywhere. In this paper, we will show the limitations of deep learning models through comparison of regression analysis and deep learning models, and present a guide for effective use of deep learning models. In addition, among various techniques used for optimization of deep learning models, data normalization and data shuffling techniques, which are widely used, are compared and evaluated based on actual data to provide guidelines for increasing the accuracy and value of deep learning models.

• Journal of the Korean Society for Precision Engineering
• /
• v.27 no.12
• /
• pp.92-98
• /
• 2010

The design of press bonding tool in LCD module equipment is a very complex and difficult task because many design able variables are involved while their effects are not known. It takes longtime experiments and much expenses to verify the effects of these design variables. However the optimization of bonding tool using OLB(outer lead bonding) and PCB Bonding is a very important problem in LCD manufacturing process, so much design efforts have been made for improving the bonding tool performance. In this paper, a reasonable and fast process which gives optimized solution under the design requirements has been presented. Both analytical and statistical methods are employed in this process. A reliable analytic model using experiment-oriented FE analysis can be obtained, in which the regression equations that predict the tool efficiency from various DOE method are found. Improvement of tool efficiency could be estimated by the regression equations using meaningful factors converged by RSM(Response Surface Method). With this process a reasonable optimized solution that meets a variety of design requirements can be easily obtained.

• Journal of ILASS-Korea
• /
• v.20 no.4
• /
• pp.247-253
• /
• 2015

Recently, the engine calibration technique has been much more complicated than that of the past engine case in order to satisfy the strict emission regulations. The current calibration method for the diesel engine which has an increasing market is both costly and time-consuming. New engine calibration method is required to develop for high-quality diesel engines with low cost and release it at the appropriate time. This study provides the optimal calibrating technique for complex engine systems using statistical modeling and numerical optimization. Firstly, it design a test plan based on Design of Experiments, a V-optimality methodology which is suitable looking for set-points, and determine the shape of test engine response. Secondly, it uses functions to make linear regression model for data analysis and optimization to fit the models of engines behavior. Finally, it generates the optimal calibrations obtained directly from empirical engine models using Grey Relational Analysis and compares the calibrations with data. This method can develop a process for systematically identifying the optimal balance of engine emissions.

• Structural Engineering and Mechanics
• /
• v.44 no.3
• /
• pp.379-403
• /
• 2012

In recent years, the need for optimal design of structures under time-history loading aroused great attention in researchers. The main problem in this field is the extremely high computational demand of time-history analyses, which may convert the solution algorithm to an illogical one. In this paper, a new framework is developed to solve the size optimization problem of steel truss structures subjected to ground motions. In order to solve this problem, the covariance matrix adaptation evolution strategy algorithm is employed for the optimization procedure, while a generalized regression neural network is utilized as a meta-model for fitness approximation. Moreover, the computational cost of time-history analysis is decreased through a wavelet analysis. Capability and efficiency of the proposed framework is investigated via two design examples, comprising of a tower truss and a footbridge truss.

• Steel and Composite Structures
• /
• v.50 no.4
• /
• pp.443-458
• /
• 2024

The construction industry, one of the biggest producers of greenhouse emissions, is under a lot of pressure as a result of growing worries about how climate change may affect local communities. Geopolymer concrete (GPC) has emerged as a feasible choice for construction materials as a result of the environmental issues connected to the manufacture of cement. The findings of this study contribute to the development of machine learning methods for estimating the properties of eco-friendly concrete, which might be used in lieu of traditional concrete to reduce CO₂ emissions in the building industry. In the present work, the compressive strength (f_c) of GPC is calculated using random forests regression (RFR) methodology where natural zeolite (NZ) and silica fume (SF) replace ground granulated blast-furnace slag (GGBFS). From the literature, a thorough set of experimental experiments on GPC samples were compiled, totaling 254 data rows. The considered RFR integrated with artificial hummingbird optimization (AHA), black widow optimization algorithm (BWOA), and chimp optimization algorithm (ChOA), abbreviated as ARFR, BRFR, and CRFR. The outcomes obtained for RFR models demonstrated satisfactory performance across all evaluation metrics in the prediction procedure. For R² metric, the CRFR model gained 0.9988 and 0.9981 in the train and test data set higher than those for BRFR (0.9982 and 0.9969), followed by ARFR (0.9971 and 0.9956). Some other error and distribution metrics depicted a roughly 50% improvement for CRFR respect to ARFR.