• International Journal of Naval Architecture and Ocean Engineering
• /
• v.9 no.1
• /
• pp.11-24
• /
• 2017

The present work focuses multi-objective optimization of blade sweep for a Wells turbine. The blade-sweep parameters at the mid and the tip sections are selected as design variables. The peak-torque coefficient and the corresponding efficiency are the objective functions, which are maximized. The numerical analysis has been carried out by solving 3D RANS equations based on k-w SST turbulence model. Nine design points are selected within a design space and the simulations are run. Based on the computational results, surrogate-based weighted average models are constructed and the population based multi-objective evolutionary algorithm gave Pareto optimal solutions. The peak-torque coefficient and the corresponding efficiency are enhanced, and the results are analysed using CFD simulations. Two extreme designs in the Pareto solutions show that the peak-torque-coefficient is increased by 28.28% and the corresponding efficiency is decreased by 13.5%. A detailed flow analysis shows the separation phenomena change the turbine performance.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.35 no.2
• /
• pp.207-212
• /
• 2011

A hydraulic breaker is an attachment installed at the end of excavator arm and is used for breaking. As per the authors' knowledge, there have been no research results on reducing the weight of the hydraulic breaker even though this weight reduction is very important for improving the performance of the excavator. In this study, we minimize the weight of the housing of the hydraulic breaker under normal operating conditions, while the maximum stress of the housing is lower than the allowable stress. A meta-model, which is generated by using the CAE results for the sampling design points determined by an orthogonal array, is used to solve the minimization problem. The weight of the housing according to the optimal design is found to be lower than the original weight by 4.8% while satisfying the constraint on the maximum stress.

• Structural Engineering and Mechanics
• /
• v.84 no.4
• /
• pp.547-560
• /
• 2022

By presetting various reinforcement diameters in topology optimization with the discrete model finite element analysis, an algorithm of bidirectional evolutionary structural optimization of multi-level reinforcement diameter is presented to obtain the optimal reinforcement topologies which describe the degree of stress of different parts. The results of a comparative study on different reinforcement feasible domain demonstrate that the more angle types of reinforcement are arranged in the initial domain, the higher utilization rate of reinforcement of the optimal topology becomes. According to the nonlinear finite element analysis of some deep beam examples, the ones designed with the optimization results have a certain advantage in ultimate bearing capacity, although their failure modes are greatly affected by the reinforcement feasible domain. Furthermore, the bearing capacity can be improved when constructional reinforcements are added in the subsequent design. However the adding would change the relative magnitude of the bearing capacity between the normal and inclined section, or the relative magnitude between the flexural and shear capacity within the inclined section, which affects the failure modes of components. Meanwhile, the adding would reduce the deformation capacity of the components as well. It is suggested that the inclined reinforcement and the constructional reinforcement should be added properly to ensure a desired ductile failure mode for components.

• Journal of the Korean Institute of Telematics and Electronics C
• /
• v.35C no.6
• /
• pp.56-66
• /
• 1998

This paper investigates a neuro-controller combined in parallel with a conventional linear controller of PID type in order to control nonminimum phase systems more efficiently. The objective is to minimize overall position errors as well as to maintain small undershooting. A costfunction is proposed with two conflict objectives. The neuro-controller is trained off-line with evolutionary programming(EP) in such a way that it becomes optimal by minimizing the given cost function through global evaluation based on desired control performance during the whole training time interval. However, it is not easy to find an optimal solution which satisfies individual objective simultaneously. With the concept of Pareto optimality and EP, we train the proposed controller more effectively and obtain a valuable set of optimal solutions. Simulation results show the efficacy of the proposed controller in a viewpoint of improvement of performance of a step response like fast settling time and small undershoot or overshoot compared with that of a conventional linear controller.

• Proceedings of the KIEE Conference
• /
• 2005.07d
• /
• pp.2888-2890
• /
• 2005

In this paper, we introduce a new structure of fuzzy-neural networks Fuzzy Set-based Polynomial Neural Networks (FSPNN). The two underlying design mechanisms of such networks involve genetic optimization and information granulation. The resulting constructs are Fuzzy Polynomial Neural Networks (FPNN) with fuzzy set-based polynomial neurons (FSPNs) regarded as their generic processing elements. First, we introduce a comprehensive design methodology (viz. a genetic optimization using Genetic Algorithms) to determine the optimal structure of the FSPNNs. This methodology hinges on the extended Group Method of Data Handling (GMDH) and fuzzy set-based rules. It concerns FSPNN-related parameters such as the number of input variables, the order of the polynomial, the number of membership functions, and a collection of a specific subset of input variables realized through the mechanism of genetic optimization. Second, the fuzzy rules used in the networks exploit the notion of information granules defined over systems variables and formed through the process of information granulation. This granulation is realized with the aid of the hard C- Means clustering (HCM). The performance of the network is quantified through experimentation in which we use a number of modeling benchmarks already experimented with in the realm of fuzzy or neurofuzzy modeling.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.6 no.1
• /
• pp.33-38
• /
• 2006

In this paper, we propose a new architecture of Fuzzy Polynomial Neural Networks(FPNN) by means of genetically optimized Fuzzy Polynomial Neuron(FPN) and discuss its comprehensive design methodology involving mechanisms of genetic optimization, especially Genetic Algorithms(GAs). The conventional FPNN developed so far are based on mechanisms of self-organization and evolutionary optimization. The design of the network exploits the extended Group Method of Data Handling(GMDH) with some essential parameters of the network being provided by the designer and kept fixed throughout the overall development process. This restriction may hamper a possibility of producing an optimal architecture of the model. The proposed FPNN gives rise to a structurally optimized network and comes with a substantial level of flexibility in comparison to the one we encounter in conventional FPNNs. It is shown that the proposed advanced genetic algorithms based Fuzzy Polynomial Neural Networks is more useful and effective than the existing models for nonlinear process. We experimented with Medical Imaging System(MIS) dataset to evaluate the performance of the proposed model.

• Journal of KIISE:Software and Applications
• /
• v.32 no.12
• /
• pp.1217-1228
• /
• 2005

Recently, since DNA computing has been widely studied for various applications, DNA sequence design which is the most basic and important step for DNA computing has been highlighted. In previous works, DNA sequence design has been formulated as a multi-objective optimization task, and solved by elitist non-dominated sorting genetic algorithm (NSGA-II). However, NSGA-II needed lots of computational time. Therefore, we use an $\varepsilon$- multiobjective evolutionarv algorithm ($\varepsilon$-MOEA) to overcome the drawbacks of NSGA-II in this paper. To compare the performance of two algorithms in detail, we apply both algorithms to the DTLZ2 benchmark function. $\varepsilon$-MOEA outperformed NSGA-II in both convergence and diversity, $70\%$ and $73\%$ respectively. Especially, $\varepsilon$-MOEA finds optimal solutions using small computational time. Based on these results, we redesign the DNA sequences generated by the previous DNA sequence design tools and the DNA sequences for the 7-travelling salesman problem (TSP). The experimental results show that $\varepsilon$-MOEA outperforms the most cases. Especially, for 7-TSP, $\varepsilon$-MOEA achieves the comparative results two tines faster while finding $22\%$ improved diversity and $92\%$ improved convergence in final solutions using the same time.

• Transactions of Materials Processing
• /
• v.23 no.8
• /
• pp.501-506
• /
• 2014

Progressive shearing with blanking dies is commonly employed to produce large quantities of tiny sheet metal electronic parts. Sheet metal pins, which are narrow and long, that are sheared with a progressive die set are often twisted. The twist in the sheet metal pins, which usually occurs in the final shearing operation, generally decreases with increasing blank holding force. The blank holding forces in all shearing operations are not the same because of different shearing positions and areas. In the current study, the optimal layout of the springs in a progressive die set to minimize the twist of the sheet metal pin is proposed. In order to find the holding force acting on the tiny narrow blanks produced with the proposed springs during the shearing process, the equivalent area method is used in the structural analysis. The shearing of the sheet-metal pin was simulated to compute the twist angle associated with the blank holding force. The constraint condition satisfying the pre-set blank holding force from the previous shearing operations was imposed. A design of experiments (DOE) was numerically implemented by analyzing the progressive die structure and by simulating the shearing process. From the meta-model created from the experimental results and by using a quadratic response surface method (PQRSM), the optimal layout of the springs was determined. The twist of sheet metal pin associated with the optimal layout of the springs found in the current study was compared with that of an existing progressive die to obtain a minimal amount of twist.

• Asian-Australasian Journal of Animal Sciences
• /
• v.25 no.6
• /
• pp.772-784
• /
• 2012

Marker-assisted gene pyramiding aims to produce individuals with superior economic traits according to the optimal breeding scheme which involves selecting a series of favorite target alleles after cross of base populations and pyramiding them into a single genotype. Inspired by the science of evolutionary computation, we used the metaphor of hill-climbing to model the dynamic behavior of gene pyramiding. In consideration of the traditional cross program of animals along with the features of animal segregating populations, four types of cross programs and two types of selection strategies for gene pyramiding are performed from a practical perspective. Two population cross for pyramiding two genes (denoted II), three population cascading cross for pyramiding three genes(denoted III), four population symmetry (denoted IIII-S) and cascading cross for pyramiding four genes (denoted IIII-C), and various schemes (denoted cross program-A-E) are designed for each cross program given different levels of initial favorite allele frequencies, base population sizes and trait heritabilities. The process of gene pyramiding breeding for various schemes are simulated and compared based on the population hamming distance, average superior genotype frequencies and average phenotypic values. By simulation, the results show that the larger base population size and the higher the initial favorite allele frequency the higher the efficiency of gene pyramiding. Parents cross order is shown to be the most important factor in a cascading cross, but has no significant influence on the symmetric cross. The results also show that genotypic selection strategy is superior to phenotypic selection in accelerating gene pyramiding. Moreover, the method and corresponding software was used to compare different cross schemes and selection strategies.

• Structural Engineering and Mechanics
• /
• v.67 no.3
• /
• pp.301-310
• /
• 2018

In this paper, a multi-objective multiparameter optimization procedure is developed by combining rigorously developed metamodels with an evolutionary search algorithm-Genetic Algorithm (GA). Response surface methodology (RSM) is used for developing the metamodels to replace the tedious finite element analyses. A nine-node isoparametric plate bending element is used for conducting the finite element simulations. Highly accurate numerical data from an author compiled FORTRAN finite element program is first used by the RSM to develop second-order mathematical relations. Four material parameters-${\frac{E_1}{E_2}}$, ${\frac{G_{12}}{E_2}}$, ${\frac{G_{23}}{E_2}}$ and ${\upsilon}_{12}$ are considered as the independent variables while simultaneously maximizing fundamental frequency, ${\lambda}_1$ and frequency separation between the $1^{st}$ two natural modes, ${\lambda}_{21}$. The optimal material combination for maximizing ${\lambda}_1$ and ${\lambda}_{21}$ is predicted by using a multi-objective GA. A general sensitivity analysis is conducted to understand the effect of each parameter on the desired response parameters.