• Journal of KIISE:Software and Applications
• /
• v.32 no.11
• /
• pp.1071-1083
• /
• 2005

By estimating probability distributions of the good solutions in the current population, some researchers try to find the optimal solution more efficiently. Particularly, finite mixtures of distributions have a very useful role in dealing with complex problems. However, it is difficult to choose the number of components in the mixture models and merge superior partial solutions represented by each component. In this paper, we propose a new continuous evolutionary optimization algorithm with distribution estimation by variational Bayesian mixtures of factor analyzers. This technique can estimate the number of mixtures automatically and combine good sub-solutions by sampling new individuals with the latent variables. In a comparison with two probabilistic model-based evolutionary algorithms, the proposed scheme achieves superior performance on the traditional benchmark function optimization. We also successfully estimate the parameters of S-system for the dynamic modeling of biochemical networks.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.10 no.4
• /
• pp.1712-1731
• /
• 2016

In this paper, we propose a novel saliency detection framework via multiple random walks (MRW) which simulate multiple agents on a graph simultaneously. In the MRW system, two agents, which represent the seeds of background and foreground, traverse the graph according to a transition matrix, and interact with each other to achieve a state of equilibrium. The proposed algorithm is divided into three steps. First, an initial segmentation is performed to partition an input image into homogeneous regions (i.e., superpixels) for saliency computation. Based on the regions of image, we construct a graph that the nodes correspond to the superpixels in the image, and the edges between neighboring nodes represent the similarities of the corresponding superpixels. Second, to generate the seeds of background, we first filter out one of the four boundaries that most unlikely belong to the background. The superpixels on each of the three remaining sides of the image will be labeled as the seeds of background. To generate the seeds of foreground, we utilize the center prior that foreground objects tend to appear near the image center. In last step, the seeds of foreground and background are treated as two different agents in multiple random walkers to complete the process of salient object detection. Experimental results on three benchmark databases demonstrate the proposed method performs well when it against the state-of-the-art methods in terms of accuracy and robustness.

• Journal of Convergence for Information Technology
• /
• v.10 no.4
• /
• pp.46-54
• /
• 2020

This study aims to develop and analyze the performance of a selective option straddle strategy based on forecasted volatility to improve the weakness of typical straddle strategy solely based on negative volatility risk premium. The KOSPI 200 option volatility is forecasted by the SVM model combined with the asymmetric volatility spillover effect. The selective straddle strategy enters option position only when the volatility is forecasted downwardly or sideways. The SVM model is trained for 2008-2014 training period and applied for 2015-2018 testing period. The suggested model showed improved performance, that is, its profit becomes higher and risk becomes lower than the benchmark strategies, and consequently typical performance index, Sharpe Ratio, increases. The suggested model gives option traders guidelines as to when they enter option position.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.40 no.2
• /
• pp.111-120
• /
• 2017

This paper proposes an improved standard genetic algorithm (GA) of making a near optimal schedule for integrated process planning and scheduling problem (IPPS) considering tool flexibility and tool related constraints. Process planning involves the selection of operations and the allocation of resources. Scheduling, meanwhile, determines the sequence order in which operations are executed on each machine. Due to the high degree of complexity, traditionally, a sequential approach has been preferred, which determines process planning firstly and then performs scheduling independently based on the results. The two sub-problems, however, are complicatedly interrelated to each other, so the IPPS tend to solve the two problems simultaneously. Although many studies for IPPS have been conducted in the past, tool flexibility and capacity constraints are rarely considered. Various meta-heuristics, especially GA, have been applied for IPPS, but the performance is yet satisfactory. To improve solution quality against computation time in GA, we adopted three methods. First, we used a random circular queue during generation of an initial population. It can provide sufficient diversity of individuals at the beginning of GA. Second, we adopted an inferior selection to choose the parents for the crossover and mutation operations. It helps to maintain exploitation capability throughout the evolution process. Third, we employed a modification of the hybrid scheduling algorithm to decode the chromosome of the individual into a schedule, which can generate an active and non-delay schedule. The experimental results show that our proposed algorithm is superior to the current best evolutionary algorithms at most benchmark problems.

• The Journal of the Acoustical Society of Korea
• /
• v.39 no.5
• /
• pp.447-453
• /
• 2020

In this paper, we propose a system to extract effective speaker representations from a speech signal using a deep learning method. Based on the fact that speech signal contains identity unrelated information such as text content, emotion, background noise, and so on, we perform a training such that the extracted features only represent speaker-related information but do not represent speaker-unrelated information. Specifically, we propose an auto-encoder based disentanglement method that outputs both speaker-related and speaker-unrelated embeddings using effective loss functions. To further improve the reconstruction performance in the decoding process, we also introduce a discriminator popularly used in Generative Adversarial Network (GAN) structure. Since improving the decoding capability is helpful for preserving speaker information and disentanglement, it results in the improvement of speaker verification performance. Experimental results demonstrate the effectiveness of our proposed method by improving Equal Error Rate (EER) on benchmark dataset, Voxceleb1.

• Journal of the Korea Computer Graphics Society
• /
• v.22 no.5
• /
• pp.1-16
• /
• 2016

In this paper, we propose and efficient curve reconstruction method based on the classical least-square fitting scheme. Specifically, given planar sample points equipped with normals, we reconstruct the objective curve as the zero set of a hierarchical implicit ZP(Zwart-Powell)-spline that can recover large holes of dataset without loosing the fine details. As regularizers, we adopted two: a Tikhonov regularizer to reduce the singularity of the linear system and a discrete Laplacian operator to smooth out the isocurves. Benchmark tests with quantitative measurements are done and our method shows much better quality than polynomial methods. Compared with the hierarchical bi-quadratic spline for datasets with holes, our method results in compatible quality but with less than 90% computational overhead.

• Journal of the Korea Concrete Institute
• /
• v.13 no.3
• /
• pp.228-236
• /
• 2001

Even though numerous researches have been performed for the prediction of thermal stresses in mass concrete structures by both analytical and experimental means, the limitations exist for both approaches. In analytical approach, the fundamental limitation is derived from the difficulty of predicting concrete properties such as modulus of elasticity, coefficient of thermal expansion, etc.. In experimental approach, there are many uncertainties related to in-situ conditions, because a majority of researches have focused on measuring thermal stresses in actual and simulated structures. In this research, an experimental device measuring thermal stresses directly in a laboratory setting is developed. The equipment is located in a temperature chamber that follows the temperature history previously obtained from temperature distribution analysis. Thermal strains are measured continuously by a strain gauge in the device and the corresponding thermal stresses are calculated simply by force equilibrium condition. For the verification of the developed device, a traditional experiment measuring thermal strains from embedded strain gauges is performed simultaneously. The results show that the thermal strain values measured by the newly developed device agree well with the results from the benchmark experiment.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.23 no.8
• /
• pp.2032-2041
• /
• 1998

In this work, we propose a novel bit allocation method that is to minimize overall distortions subject ot the bit rate constraint. We partition the original bitallocation problem into 'macroblock level bit allocation' problems that can be solved by conventional Lagrangian mutiplier methods and a 'frame level bit allocation' problem. To tackle the frame level problem, 'two-phase optimization' algorithm is used with iter-frame dependency model. While the existing approaches are almost impossible to find the macroblock-unit result for the moving picture coding system due to high computational complexity, the proposed algorithm can drastically reduce the computational loads by the problem partitioning and can obtain the result close to the optimal solution. Because the optimally allocated results can be used as a benchmark for bit allocation methods, the upper performance limit, or a basis for approximation method development, we expect that the proposed algorithm can be very useful for the bit allocation related works.

• Steel and Composite Structures
• /
• v.33 no.5
• /
• pp.717-727
• /
• 2019

This paper is motivated by the lack of studies in the technical literature concerning to vibration analysis of a single-layered graphene sheet (SLGS) with corner cutout based on the nonlocal elasticity model framework of classical Kirchhoff thin plate. An isogeometric analysis (IGA) based upon non-uniform rational B-spline (NURBS) is employed for approximation of the L-shape SLGS deflection field. Trimming technique is employed to create the cutout in geometry of L-shape plate. The L-shape plate is assumed to be Free (F) in the straight edges of cutout while any arbitrary boundary conditions are applied to the other four straight edges including Simply supported (S), Clamped (C) and Free (F). The Numerical studies are carried out to express the influences of the nonlocal parameter, cutout dimensions, boundary conditions and mode numbers on the variations of the natural frequencies of SLGS. It is precisely shown that these parameters have considerable effects on the free vibration behavior of the system. In addition, numerical results are validated and compared with those achieved using other analysis, where an excellent agreement is found. The effectiveness and the accuracy of the present IGA approach have been demonstrated and it is shown that the IGA is efficient, robust and accurate in terms of nanoplate problems. This study serves as a benchmark for assessing the validity of numerical methods used to analyze the single-layered graphene sheet with corner cutout.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.46 no.6
• /
• pp.58-64
• /
• 2009

For the success of the reconfigurable computing, the algorithm for mapping operations onto coarse-grained reconfigurable architecture is very important. This paper proposes a resource-aware mapping system for the coarse-grained reconfigurable architecture and its own underlying heuristic algorithm. The operation assignment and the routing path allocation are simultaneously performed with a cycle-accurate time-exclusive resource model. The proposed algorithm minimizes the communication resource usage and the global memory access with the list scheduling heuristic. The operation to be mapped are prioritized with general properties of data flow. The evaluations of the proposed algorithm show that the performance is significantly enhanced in several benchmark applications.