• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.30 no.6 s.249
• /
• pp.691-697
• /
• 2006

Metamodel, model of model, has been widely used to improve an efficiency of optimization process in engineering fields. However, global metamodels of constraints in a constrained optimization problem are required good accuracy around neighborhood of optimum point. To satisfy this requirement, more sampling points must be located around the boundary and inside of feasible region. Therefore, a new sampling strategy that is capable of identifying feasible domain should be applied to select sampling points for metamodels of constraints. In this research, we suggeste sequential feasible domain sampling that can locate sampling points likely within feasible domain by using penalty function method. To validate the excellence of feasible domain sampling, we compare the optimum results from the proposed method with those form conventional global space-filling sampling for a variety of optimization problems. The advantages of the feasible domain sampling are discussed further.

• Communications for Statistical Applications and Methods
• /
• v.6 no.1
• /
• pp.143-150
• /
• 1999

Ko(1998) proposed a procedure to enhance the efficiency of double sampling plans by allowing second-stage sample size and critical region to depend on first-stage evidence using constraint optimization approaches. In this study further developments of such plans by incorporating several practically possible researcher's aims into the optimization are considered. Comparisons are made with the optimal ordinary double sampling plan and also among them It is observed that it is to some extent possible to match the details of the optimization to certain qualitative methodological aims.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.10 no.5
• /
• pp.117-124
• /
• 2011

Tire's performance plays important roles in improving vehicle's performances. Tire makers carry out a lot of research to improve tire's performance. They are making effort to meet multi purposes using various optimization methods. Recently, the tire makers perform the shape optimization using approximation models, which are surrogate models obtained by statistical method. Generally, the reason why we increase sampling points during optimization process, is to get more reliable approximation models, but the more we adopt sampling points, the more we need time to test. So it is important to select approximation model and proper number of sampling points to balance between reliability and time consuming. In this research, we studied to compare two kind cases for a approximation construction. First, we compare RSM and Kriging which are Curve Fitting Method and Interpolation Method, respectively. Second, we construct approximation models using three different number of sampling points. And then, we recommend proper approximation model and orthogonal array adopt tire's design optimization.

• Proceedings of the KIEE Conference
• /
• 2007.04c
• /
• pp.162-164
• /
• 2007

This paper presents a "window-zoom-out" optimization strategy with relatively fewer sampling data. In this method, an optimal Latin hypercube sampling experiment based on multi-objective Pareto optimization is developed to obtain the sampling data. The response surface method with multiquadric radial basis function combined with (1+$\lambda$) evolution strategy is used to find the global optimal point. The proposed method is verified with numerical experiments.

• Korean Chemical Engineering Research
• /
• v.52 no.5
• /
• pp.564-573
• /
• 2014

Economic optimization of cumene manufacturing process to produce cumene from benzene and propylene was studied. The chosen objective function was the operational profit per year that subtracted capital cost, utility cost, and reactants cost from product revenue and other benefit. The number of design variables of the optimization are 6. Matlab connected to and controlled Unisim Design to calculate operational profit with the given design variables. As the first step of the optimization, design variable points was sampled and operational profit was calculated by using Unisim Design. By using the sampled data, the estimation model to calculate the operational profit was constructed, and the optimization was performed on the estimation model. This study compared second order polynomial and support vector regression as the estimation method. As the sampling method, central composite design was compared with Hammersley sequence sampling. The optimization results showed that support vector regression and Hammersley sequence sampling were superior than second order polynomial and central composite design, respectively. The optimized operational profit was 17.96 MM$ per year, which was 12% higher than 16.04 MM$ of base case.

• Proceedings of the KIEE Conference
• /
• 2008.04c
• /
• pp.65-67
• /
• 2008

An adaptive response surface method with Latin Hypercube sampling strategy is employed to optimize a magnet pole shape of large scale BLDC motor to minimize the cogging torque. The proposed algorithm consists of the multi-objective Pareto optimization and ($1+{\lambda}$) evolution strategy to find the global optimal points with relatively fewer sampling data. In the adaptive RSM, an adaptive sampling point insertion method is developed utilizing the design sensitivities computed by using finite element method to set a reasonable response surface with a relatively small number of sampling points. The developed algorithm is applied to the shape optimization of PM poles for 6MW BLDC motor.

• Proceedings of the KIEE Conference
• /
• 2008.07a
• /
• pp.701-702
• /
• 2008

An adaptive response surface method with Latin Hypercube sampling strategy is employed to optimize a magnet pole shape of large scale BLDC motor to minimize the cogging torque. The proposed algorithm consists of the multi-objective Pareto optimization and (1+${\lambda}$) evolution strategy to find the global optimal points with relatively fewer sampling data. In the adaptive RSM, an adaptive sampling point insertion method is developed utilizing the design sensitivities computed by using finite element method to get a reasonable response surface with a relatively small number of sampling points. The developed algorithm is applied to the shape optimization of PM poles for 6 MW BLDC motor, and the cogging torque is reduced to 19% of the initial one.

• International Journal of Naval Architecture and Ocean Engineering
• /
• v.10 no.6
• /
• pp.730-740
• /
• 2018

Hydrofoil is the direct component to generate thrust for underwater glider. It is significant to improve propulsion efficiency of hydrofoil. This study optimizes the shape of a hydrofoil using Free-Form Deformation (FFD) parametric approach and Surrogate-based Optimization (SBO) algorithm. FFD approach performs a volume outside the hydrofoil and the position changes of control points in the volume parameterize hydrofoil's geometric shape. SBO with adaptive parallel sampling method is regarded as a promising approach for CFD-based optimization. Combination of existing sampling methods is being widely used recently. This paper chooses several well-known methods for combination. Investigations are implemented to figure out how many and which methods should be included and the best combination strategy is provided. As the hydrofoil can be stretched from airfoil, the optimizations are carried out on a 2D airfoil and a 3D hydrofoil, respectively. The lift-drag ratios are compared among optimized and original hydrofoils. Results show that both lift-drag-ratios of optimized hydrofoils improve more than 90%. Besides, this paper preliminarily explores the optimization of hydrofoil with root-tip-ratio. Results show that optimizing 3D hydrofoil directly achieves slightly better results than 2D airfoil.

• International Journal of Fluid Machinery and Systems
• /
• v.10 no.3
• /
• pp.240-253
• /
• 2017

Latin hypercube sampling is widely used design-of-experiment technique to select design points for simulation which are then used to construct a surrogate model. The exploration/exploitation properties of surrogate models depend on the size and distribution of design points in the chosen design space. The present study aimed at evaluating the performance characteristics of various surrogate models depending on the Latin hypercube sampling (LHS) procedure (sample size and spatial distribution) for a diverse set of optimization problems. The analysis was carried out for two types of problems: (1) thermal-fluid design problems (optimizations of convergent-divergent micromixer coupled with pulsatile flow and boot-shaped ribs), and (2) analytical test functions (six-hump camel back, Branin-Hoo, Hartman 3, and Hartman 6 functions). The three surrogate models, namely, response surface approximation, Kriging, and radial basis neural networks were tested. The important findings are illustrated using Box-plots. The surrogate models were analyzed in terms of global exploration (accuracy over the domain space) and local exploitation (ease of finding the global optimum point). Radial basis neural networks showed the best overall performance in global exploration characteristics as well as tendency to find the approximate optimal solution for the majority of tested problems. To build a surrogate model, it is recommended to use an initial sample size equal to 15 times the number of design variables. The study will provide useful guidelines on the effect of initial sample size and distribution on surrogate construction and subsequent optimization using LHS sampling plan.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.8 no.6
• /
• pp.2070-2086
• /
• 2014

As a kind of personal lifelog data, activity data have been considered as one of the most compelling information to understand the user's habits and to calibrate diagnoses. In this paper, we proposed a robust algorithm to sampling rates for human activity recognition, which identifies a user's activity using accelerations from a triaxial accelerometer in a smartphone. Although a high sampling rate is required for high accuracy, it is not desirable for actual smartphone usage, battery consumption, or storage occupancy. Activity recognitions with well-known algorithms, including MLP, C4.5, or SVM, suffer from a loss of accuracy when a sampling rate of accelerometers decreases. Thus, we start from particle swarm optimization (PSO), which has relatively better tolerance to declines in sampling rates, and we propose PSO with an adaptive boundary correction (ABC) approach. PSO with ABC is tolerant of various sampling rate in that it identifies all data by adjusting the classification boundaries of each activity. The experimental results show that PSO with ABC has better tolerance to changes of sampling rates of an accelerometer than PSO without ABC and other methods. In particular, PSO with ABC is 6%, 25%, and 35% better than PSO without ABC for sitting, standing, and walking, respectively, at a sampling period of 32 seconds. PSO with ABC is the only algorithm that guarantees at least 80% accuracy for every activity at a sampling period of smaller than or equal to 8 seconds.