• Korean Journal of Computational Design and Engineering
• /
• v.5 no.2
• /
• pp.127-135
• /
• 2000

Emergent computing paradigms such as genetic algorithms have found increased use in problems in engineering design. These computational tools have been shown to be applicable in the solution of generically difficult design optimization problems characterized by nonconvexities in the design space and the presence of discrete and integer design variables. Another aspect of these computational paradigms that have been lumped under the bread subject category of soft computing, is the domain of artificial intelligence, knowledge-based expert system, and machine learning. The paper explores a machine learning paradigm referred to as teaming classifier systems to construct the high-quality global function approximations between the design variables and a response function for subsequent use in design optimization. A classifier system is a machine teaming system which learns syntactically simple string rules, called classifiers for guiding the system's performance in an arbitrary environment. The capability of a learning classifier system facilitates the adaptive selection of the optimal number of training data according to the noise and multimodality in the design space of interest. The present study used the polynomial based response surface as global function approximation tools and showed its effectiveness in the improvement on the approximation performance.

• Journal of Magnetics
• /
• v.20 no.2
• /
• pp.193-200
• /
• 2015

Direct-acting solenoid valves are used in the automotive industry due to their simple structure and quick response in controlling the flow of fluid. We performed an optimization study of response time in order to improve the dynamic performance of a direct-acting solenoid valve. For the optimal design process, we used the commercial optimization software PIAnO, which provides various tools for efficient optimization including design of experiments (DOE), approximation techniques, and a design optimization algorithm. 35 sampling points of computational experiments are performed to find the optimum values of the design variables. In all cases, ANSYS Maxwell electromagnetic analysis software was used to model the electromagnetic dynamics. An approximate model generated from the electromagnetic analysis was estimated and used for the optimization. The best optimization model was selected using the verified approximation model called the Kriging model, and an optimization algorithm called the progressive quadratic response surface method (PQRSM).

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.29 no.3 s.234
• /
• pp.369-386
• /
• 2005

Various aspects of structural optimization techniques under transient loads are extensively reviewed. The main themes of the paper are treatment of time dependent constraints, calculation of design sensitivity, and approximation. Each subject is reviewed with the corresponding papers that have been published since 1970s. The treatment of time dependent constraints in both the direct method and the transformation method is discussed. Two ways of calculating design sensitivity of a structure under transient loads are discussed - direct differentiation method and adjoint variable method. The approximation concept mainly focuses on re- sponse surface method in crashworthiness and local approximation with the intermediate variable Especially, as an approximated optimization technique, Equivalent Static Load method which takes advantage of the well-established static response optimization technique is introduced. And as an application area of dynamic response optimization technique, the structural optimization in flexible multibody dynamic systems is re- viewed in the viewpoint of the above three themes

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.33 no.2
• /
• pp.32-38
• /
• 2005

In this research, a procedure to improve the feasibility of design space is proposed by an approximation model. The Chebyshev Inequality is used as the criterion of modification of design space. This procedure is applied to the aero-elastic transonic wing design problem and the feasibility of the design space is greatly improved. Also the optimization results are improved by appling this procedure. That is, the probability to satisfy all imposed constraints is increased and the better design points are included in design space after this procedure. And the use of both a second-order response surface model and the Kriging model is investigated and compared in accuracy, efficiency, and robustness as approximation models in this procedure for different sampling methods. As a result, the second-order response surface model is more appropriate for our application than the Kriging model, because it is linear enough to be fitted well by the response surface model.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2004.10a
• /
• pp.700-703
• /
• 2004

A reliability analysis and design procedure based on the design of experiment (DOE) is combined with the response surface method (RSM) for numerical efficiency. The procedure established is based on a 3$^n$ full factorial DOE for numerical quadrature using explicit formula of optimum levels and weights derived for general distributions. The full factorial moment method (FFMM) shows good performance in terms of accuracy and ability to treat non-normally distributed random variables. But, the FFMM becomes very inefficient because the number of function evaluation required increases exponentially as the number of random variables considered increases. To enhance the efficiency, the response surface moment method (RSMM) is proposed. In RSMM, experiments only with high probability are conducted and the rest of data are complemented by a quadratic response surface approximation without mixed terms. The response surface is updated by conducting experiments one by one until the value of failure probability is converged. It is calculated using the Pearson system and the four statistical moments obtained from the experimental data. A measure for checking the relative importance of an experimental point is proposed and named as influence index. During the update of response surface, mixed terms can be added into the formulation.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.16 no.3
• /
• pp.30-37
• /
• 2017

When a wheel profile of a train-tram is designed, both train and tram tracks should be considered. This study designed a wheel profile that enables high-speed driving(200km/h) on the train track and low speed driving on the tram track with multiple sharp curves. The study used the approximation optimization method to reduce cost and time, used the sequential quadratic programming method as the optimized algorithm, and the central composite design and response surface method as an approximate model. The optimized wheel shape based on this approximation optimization method reduced wear of the initial wheel showed a better performance in terms of derailment and lateral force.

• Proceedings of the KIEE Conference
• /
• 2003.07b
• /
• pp.994-996
• /
• 2003

The Response Surface Methodology is frequently used for building an approximation model. However, its approximation errors often occur in engineering problem, because of the use of the Least Square Method. Therefore, this paper introduces the Moving Least Square Method to obtain the more accurate Response Surface Model, and then the detent force of a Permanent Magnet Linear Synchronous Motor is applied to verify the accuracy of the introduced method.

• International Journal of Automotive Technology
• /
• v.8 no.1
• /
• pp.93-102
• /
• 2007

The frontal crash optimization of an engine room member using the response surface method was studied. The engine room member is composed of the front side member and the sub-frame. The thicknesses of the panels on the front side member and the sub-frame were selected as the design variables. The purpose of the optimization was to reduce the weight of the structure, under the constraint that the objective quantity of crash energy is absorbed. The response surface method was used to approximate the crash behavior in mathematical form for optimization procedure. To research the effect of the regression method, two different methodologies were used in constructing the response surface model, the least square method and the moving least square method. The optimum with the two methods was verified by the simulation result. The precision of the surrogate model affected the optimal design. The moving least square method showed better approximation than the least square method. In addition to the deterministic optimization, the reliability-based design optimization using the response surface method was executed to examine the effect of uncertainties in design variables. The requirement for reliability made the optimal structure be heavier than the result of the deterministic optimization. Compared with the deterministic optimum, the optimal design using the reliability-based design optimization showed higher crash energy absorption and little probability of failure in achieving the objective.

• Korean Journal of Metals and Materials
• /
• v.46 no.7
• /
• pp.464-470
• /
• 2008

Metallic nanostructures have great potential for bio-chemical sensor applications due to the excitation of localized surface plasmon and its sensitive response to environmental change. Unlike the commonly explored absorption-based sensing, the optical scattering provides single particle detection scheme. For the localized surface plasmon resonance spectroscopy, the metallic nanostructures with controlled shape and size have been usually fabricated on adhesion-layer pre-coated transparent glass substrates. In this study, we calculated the optical scattering properties of plasmonic Au nanodisc using a discrete dipole approximation method and analyzed the effect of adhesion layer on them. Our result also indicates that there is a trade-off between the surface plasmon damping and the capability of supporting nanostructures in determining the optimal thickness of adhesion layer. Marginal thickness of Ti adhesion layer for supporting Au nanostructures fabricated on a silica glass substrate was experimentally analyzed by an adhesion strength test using a nano-indentation technique.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.28 no.6A
• /
• pp.835-842
• /
• 2008

The response surface method (RSM) is widely adopted for the structural reliability analysis because of its numerical efficiency. However, the RSM is still time consuming for large-scale applications and sometimes shows large errors in the calculation of sensitivity of reliability index with respect to random variables. Therefore, this study proposes a new RSM in which moving least squares (MLS) approximation is applied. Least squares approximation generally used in the common RSM gives equal weight to the coefficients of the response surface function (RSF). On the other hand, The MLS approximation gives higher weight to the experimental points closer to the design point, which yields the RSF more similar to the limit state at the design point. In the procedure of the proposed method, a linear RSF is constructed initially and then a quadratic RSF is formed using the axial experimental points selected from the reduced region where the design point is likely to exist. The RSF is updated successively by adding one more experimental point to the previously sampled experimental points. In order to demonstrate the effectiveness of the proposed method, mathematical problems and ten-bar truss are considered as numerical examples. As a result, the proposed method shows better accuracy and computational efficiency than the common RSM.