• Journal of the Korean Data and Information Science Society
• /
• v.16 no.2
• /
• pp.255-261
• /
• 2005

Many designed experiments require the simultaneous optimization of multiple responses. In this paper, we propose how to simultaneously optimize multiple responses for robust design when data are collected from a combined array. The proposed method is based on the quadratic loss function. An example is illustrated to show the proposed method.

• The KIPS Transactions:PartB
• /
• v.14B no.5
• /
• pp.377-382
• /
• 2007

This paper proposes a training algorithm for global optimization of the parameters of radial basis function networks. Since conventional training algorithms usually perform only local optimization, the performance of the network is limited and the final network significantly depends on the initial network parameters. The proposed hybrid simulated annealing algorithm performs global optimization of the network parameters by combining global search capability of simulated annealing and local optimization capability of gradient-based algorithms. Via experiments for function approximation problems, we demonstrate that the proposed algorithm can find networks showing better training and test performance and reduce effects of the initial network parameters on the final results.

• International Journal of Computer Science & Network Security
• /
• v.21 no.4
• /
• pp.33-40
• /
• 2021

The traveling salesman problem (TSP) is one of the well-known and extensively studied NPC problems in combinatorial optimization. To solve it effectively and efficiently, various optimization algorithms have been developed by scientists and researchers. However, most optimization algorithms are designed based on the concept of improving route in the iterative improvement process so that the optimal solution can be finally found. In contrast, there have been relatively few algorithms to find the optimal solution using route construction mechanism. In this paper, we propose a route construction optimization algorithm to solve the symmetric TSP with the help of ratio value. The proposed algorithm starts with a set of sub-routes consisting of three cities, and then each good sub-route is enhanced step by step on both ends until feasible routes are formed. Before each subsequent expansion, a ratio value is adopted such that the good routes are retained. The experiments are conducted on a collection of benchmark symmetric TSP datasets to evaluate the algorithm. The experimental results demonstrate that the proposed algorithm produces the best-known optimal results in some cases, and performs better than some other route construction optimization algorithms in many symmetric TSP datasets.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.26 no.1
• /
• pp.87-92
• /
• 2016

In this paper, optimization technique such as particle swarm optimization was used to optimize the parameters of fuzzy Extreme Learning Machine. While the learning speed of conventional neural networks is very slow, that of Extreme Learning Machine is very fast. Fuzzy Extreme Learning Machine is composed of the Extreme Learning Machine with very fast learning speed and fuzzy logic which can represent the linguistic information of the field experts. The general sigmoid function is used for the activation function of Extreme Learning Machine. However, the activation function of Fuzzy Extreme Learning Machine is the membership function which is defined in the procedure of fuzzy C-Means clustering algorithm. We optimize the parameters of the membership functions by using optimization technique such as Particle Swarm Optimization. In order to validate the classification capability of the proposed classifier, we make several experiments with the various machine learning datas.

• Nuclear Engineering and Technology
• /
• v.51 no.5
• /
• pp.1279-1288
• /
• 2019

The vortex-type fluidic diode (FD) is a key safety component for inherent safety in various advanced reactors such as the sodium fast reactor (SFR) and the molten salt reactor (MSR). In this study, topology optimization is conducted to optimize the design of the vortex-type fluidic diode. The optimization domain is simplified to 2-dimensional geometry for a tangential port and chamber. As a result, a design with a circular chamber and a restrictor at the tangential port is obtained. To verify the new design, experimental study and computational fluid dynamics (CFD) analysis were conducted for inlet Reynolds numbers between 2000 and 6000. However, the results show that the performance of the new design is no better than the original reference design. To analyze the cause of this result, detailed analysis is performed on the velocity and pressure field using flow visualization experiments and 3-D CFD analysis. The results show that the discrepancy between the optimization results in 2-D and the experimental results in 3-D originated from exclusion of an important pressure loss contributor in the optimization process. This study also concludes that the junction design of the axial port and chamber offers potential for improvement of fluidic diode performance.

• KSBB Journal
• /
• v.29 no.3
• /
• pp.165-178
• /
• 2014

Statistical medium optimization has been carried out for the production of succinic acid in anaerobic fermentations of Actinobacillus succinogenes. Succinic acid utilized as a precursor of many industrially important chemicals is a fourcarbon dicarboxylic acid, biosynthesized as one of the fermentation products of anaerobic metabolism by A. succinogenes. Through OFAT (one factor at a time) experiments, corn steep liquor (CSL), a very cheap agricultural byproduct, was found to have significant effects on enhanced production of succinic acid, when supplemented along with yeast extract. Hence, using these factors including glucose as a carbon/energy source, interactive effects were investigated through $2^n$ full factorial design (FFD) experiments, showing that the concentration of each component (i.e., glucose, yeast extract and CSL) should be higher. Further statistical experiments were conducted along the steepest ascent path, followed by response surface method (RSM) in order to find out optimal concentrations of each constituent. Consequently, optimized concentrations of glucose, yeast extract and CSL were observed to be 180 g/L, 15.08 g/L and 20.75 g/L respectively (10 g/L of $NaHCO_3$ and 100 g/L of $MgCO_3$ to be supplemented as bicarbonate suppliers), with the estimated production level of succinic acid to be 92.9 g/L (about 3.5 fold higher productivity as compared to the initial medium). Notably, the RSM-estimated production level was almost similar to the amount of succinic acid (92.9 g/L vs. 89.1 g/L) produced through the actual fermentation process performed using the statistically optimized production medium.

• Journal of the Computational Structural Engineering Institute of Korea
• /
• v.16 no.4
• /
• pp.345-351
• /
• 2003

Due to increasing legal and market demands for safety in the automotive design process, the design of integrated seat is important more and mote because it should satisfy the conflict between stronger and lower weight for safety and environmental demands. In this study for crash simulations, the numerical simulations have been carried out using the explicit finite element program LS-Dyna according to the FMVSS 210 standard for safety test of seat. Since crash simulations are very time-consuming and a series of simulations that does not lead to a better result is very costly, the optimization method must be both efficient and reliable. As a result of that, statistical approaches such as design of experiments and response surface model have been successfully implemented to reduce time-consuming LS-Dyna simulations and optimize the safety and environmental demands together with nonlinear optimization algorithm. Design of experiments is used lot exploring the design space of maximum displacement and total weight and for building response surface models in order to minimize the maximum displacement and total weight of integrated seat.

• The KIPS Transactions:PartB
• /
• v.18B no.1
• /
• pp.39-44
• /
• 2011

In this paper, we propose a method to improve the Firefly Algorithm(FA) introduced by Xin-She Yang, recently. We design and analyze the improved firefly algorithm based on the heuristic. We compare the FA with the Particle Swarm Optimization (PSO) which the problem domain is similar with the FA in terms of accuracy, algorithm convergence time, the motion of each particle. The compare experiments show that the accuracy of FA is not worse than PSO's, but the convergence time of FA is slower than PSO's. In this paper, we consider intuitive reasons of slow convergence time problem of FA, and propose the improved version of FA using a partial mutation heuristic based on the consideration. The experiments using benchmark functions show the accuracy and convergence time of the improved FA are better than them of PSO and original FA.

• MALSORI
• /
• no.62
• /
• pp.149-161
• /
• 2007

In this paper we proposed a modified reliability function for improving bimodal speaker identification(BSI) performance. The convectional reliability function, used by N. Fox[1], is extended by introducing an optimization factor. We evaluated the proposed method in BSI domain. A BSI system was implemented based on GMM and it was tested using VidTIMIT database. Through speaker identification experiments we verified the usefulness of our proposed method. The experiments showed the improved performance, i.e., the reduction of error rate by 39%.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.16 no.2
• /
• pp.7-15
• /
• 2017

Recently, an electricity sensor system has been installed and operated to prevent failures and accidents by identifying whether a transformer is normal in advance of failure. This electricity sensor system is able to both measure and monitor the transformer's power and voltage remotely and send information to a manager when unusual operation is discovered. However, a battery is required to operate power detection devices, and battery systems need ongoing management such as regular replacement. In addition, at a maintenance cost, occasional human resources and worker safety problems arise. Accordingly, we apply a linear electromagnetic generator using vibration energy from a transformer for an electric sensor system's drive in this research and we conduct optimal design to maximize the linear electromagnetic generator's power. We consider design variables using the provided design method from Process Integration, Automation, and Optimization (PIAnO), which is common tool from process integration and design optimization (PIDO). In addition, we analyze the experiment point from the design of the experiments using "MAXWELL," which is a common electromagnet analysis program. We then create an approximate model and conduct accuracy verification. Finally, we determine the optimal model that generates the maximum power using the proven approximate kriging model and evolutionary optimization algorithm, which we then confirm via simulation.