• Journal of Korea Water Resources Association
• /
• v.55 no.11
• /
• pp.903-911
• /
• 2022

Groundwater, one of the resources for supplying water, fluctuates in water level due to various natural factors. Recently, research has been conducted to predict fluctuations in groundwater levels using Artificial Neural Network (ANN). Previously, among operators in ANN, Gradient Descent (GD)-based Optimizers were used as Optimizer that affect learning. GD-based Optimizers have disadvantages of initial correlation dependence and absence of solution comparison and storage structure. This study developed Gradient Descent combined with Harmony Search (GDHS), a new Optimizer that combined GD and Harmony Search (HS) to improve the shortcomings of GD-based Optimizers. To evaluate the performance of GDHS, groundwater level at Icheon Yullhyeon observation station were learned and predicted using Multi Layer Perceptron (MLP). Mean Squared Error (MSE) and Mean Absolute Error (MAE) were used to compare the performance of MLP using GD and GDHS. Comparing the learning results, GDHS had lower maximum, minimum, average and Standard Deviation (SD) of MSE than GD. Comparing the prediction results, GDHS was evaluated to have a lower error in all of the evaluation index than GD.

• Structural Engineering and Mechanics
• /
• v.63 no.3
• /
• pp.269-280
• /
• 2017

One major concern that has occupied the mind of the designers is a structural failure as result of stress concentration in the geometrical discontinuities. Understanding the effective parameters contribute to stress concentration and proper selection of these parameters enables the designer get to a reliable design. In the analysis of perforated isotropic and orthotropic plates, the effective parameters on stress distribution around holes include load angle, curvature radius of the corner of the hole, hole orientation and fiber angle for orthotropic materials. This present paper tries to examine the possible effects of these parameters on stress analysis of infinite perforated plates with central quasi-square hole applying grey wolf optimizer (GWO) inspired by the particular leadership hierarchy and hunting behavior of grey wolves in nature, and also the present study tries to introduce general optimum parameters in order to achieve the minimum amount of stress concentration around this type of hole on isotropic and orthotropic plates. The advantages of grey wolf optimizer are stout, flexible, simple, and easy to be enforced. The used analytical solution is the expansion of Lekhnitskii's solution method. Lekhnitskii applied this method for the stress analysis of anisotropic plates containing circular and elliptical holes. Finite element numerical solution is employed to examine the results of present analytical solution. Results represent that by selecting the aforementioned parameters properly, fewer amounts of stress could be achieved around the hole leading to an increase in load-bearing capacity of the structure.

• Structural Engineering and Mechanics
• /
• v.61 no.3
• /
• pp.359-370
• /
• 2017

This study presents a particle swarm optimization algorithm integrated with weighted particle concept and improved fly-back technique. The rationale behind this integration is to utilize the affirmative properties of these new terms to improve the search capability of the standard particle swarm optimizer. Improved fly-back technique introduced in this study can be a proper alternative for widely used penalty functions to handle existing constraints. This technique emphasizes the role of the weighted particle on escaping from trapping into local optimum(s) by utilizing a recursive procedure. On the other hand, it guaranties the feasibility of the final solution by rejecting infeasible solutions throughout the optimization process. Additionally, in contrast with penalty method, the improved fly-back technique does not contain any adjustable terms, thus it does not inflict any extra ad hoc parameters to the main optimizer algorithm. The improved fly-back approach, as independent unit, can easily be integrated with other optimizers to handle the constraints. Consequently, to evaluate the performance of the proposed method on solving the truss weight minimization problems with discrete variables, several benchmark examples taken from the technical literature are examined using the presented method. The results obtained are comparatively reported through proper graphs and tables. Based on the results acquired in this study, it can be stated that the proposed method (integrated particle swarm optimizer, iPSO) is competitive with other metaheuristic algorithms in solving this class of truss optimization problems.

• International Journal of Internet, Broadcasting and Communication
• /
• v.11 no.4
• /
• pp.31-36
• /
• 2019

Unsupervised neural networks have not caught enough attention until Generative Adversarial Network (GAN) was proposed. By using both the generator and discriminator networks, GAN can extract the main characteristic of the original dataset and produce new data with similarlatent statistics. However, researchers understand fully that training GAN is not easy because of its unstable condition. The discriminator usually performs too good when helping the generator to learn statistics of the training datasets. Thus, the generated data is not compelling. Various research have focused on how to improve the stability and classification accuracy of GAN. However, few studies delve into how to improve the training efficiency and to save training time. In this paper, we propose a novel optimizer, named FAST-ADAM, which integrates the Lookahead to ADAM optimizer to train the generator of a semi-supervised generative adversarial network (SSGAN). We experiment to assess the feasibility and performance of our optimizer using Canadian Institute For Advanced Research - 10 (CIFAR-10) benchmark dataset. From the experiment results, we show that FAST-ADAM can help the generator to reach convergence faster than the original ADAM while maintaining comparable training accuracy results.

• Journal of KIISE:Software and Applications
• /
• v.30 no.1_2
• /
• pp.73-80
• /
• 2003

The Java programming language is designed for developing effective applications in a heterogeneous network environment. Major problem in Java is its performance. many attractive features of Java make the development of software easy, but also make it expensive to support ; applications written in Java are often much slower than their counterparts written in C or C++. To use Java`s attractive features without the performance penalty, sophisticated optimizations and runtime systems are required. Optimising Java bytecode has several advantages. First, the bytecode is independent of any compiler that is used to generate it. Second, the bytecode optimization can be performed as a pre=pass to Just-In-Time(JIT) compilation. Many attractive features of Java make the development of software easy, but also make it expensive to support. The goal of this work is to develop automatic construction of code optimizer for Java bytecode. We`ve designed and implemented a Bytecode Optimizer that performs the peephole optimization, bytecode-specific optimization, and method-inlining techniques. Using the Classfile optimizer, we see up to 9% improvement in speed and about 20% size reduction in Java class files, when compared to average code using the interpreter alone.

• The KIPS Transactions:PartA
• /
• v.14A no.3 s.107
• /
• pp.167-172
• /
• 2007

VM(Virtual Machine) can be considered as a software processor which interprets the machine code. Also, it is considered as a conceptional computer that consists of logical system configuration. But, the execution speed of VM system is much slower than that of a real processor system. So, it is very important to optimize the code for virtual machine to enhance the execution time. Especially the optimizer for a virtual machine code on embedded devices requires the highly efficient performance to the ordinary optimizer in the respect to the optimized ratio about cost. Fundamentally, functions and basic blocks which influence the execution time of virtual machine is found, and then an optimization for them nay get the high efficiency. In this paper, we designed and implemented the optimizer for the virtual(or abstract) machine code(VMC) using profiling. Firstly, we defined the profiling information which is necessary to the optimization of VMC. The information can be obtained from dynamically executing the machine code. And we implemented VMC optimizer using the profiling information. In our implementation, the VMC is SIL(Standard Intermediate Language) that is an intermediate code of EVM(Embedded Virtual Machine). Also, we tried a benchmark test for the VMC optimizer and obtained reasonable results.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.23 no.11
• /
• pp.1357-1363
• /
• 2019

In this paper, we compares the performance of the gredient descent optimizers of the Faster Region-based Convolutional Neural Network (R-CNN) model for the chromosome object detection in digital images composed of human metaphase chromosomes. In faster R-CNN, the gradient descent optimizer is used to minimize the objective function of the region proposal network (RPN) module and the classification score and bounding box regression blocks. The gradient descent optimizer. Through performance comparisons among these four gradient descent optimizers in our experiments, we found that the Adamax optimizer could achieve the mean average precision (mAP) of about 52% when considering faster R-CNN with a base network, VGG16. In case of faster R-CNN with a base network, ResNet50, the Adadelta optimizer could achieve the mAP of about 58%.

• Korean Journal of Air-Conditioning and Refrigeration Engineering
• /
• v.18 no.8
• /
• pp.635-640
• /
• 2006

An optimizer has been applied for the optimal design of pipe diameters in the pipe flow network problems. Pipe network flow analysis, which is developed separately, is performed within the interface for the optimization algorithm. A pipe network is chosen for the test, and optimizer GenOpt is applied with Holder-Mead-O'Niell's simplex algorithm after solving the network flow problem by the Newton-Raphson method. As a result, optimally do-signed pipe diameters are successfully obtained which minimize the total design cost. Design cost of pipe flow network can be considered as the sum of pipe installation cost and pump operation cost. In this study, a practical and efficient solution method for the pipe network optimization is presented. Test system is solved for the demonstration of the present optimization technique.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.57 no.12
• /
• pp.2285-2290
• /
• 2008

Even though efficiency of coal-fired power plant is proportional to operating temperature, increasement of operating temperature is limited by a technological level of each power plant component. It is an alternative plan to increase operating pressure up to ultra super critical point for efficiency enhancement. It is difficult to control process of power plant in ultra super critical point because that point has highly nonlinear characteristics. In this paper, new control logic, Unit Response Optimizer Controller(URO Controller) which is based on Fuzzy logic and Model Predictive Controller, is introduced for better performance. Then its performance is tested and analyzed with design guideline.

• Proceedings of the Korean Information Science Society Conference
• /
• 2001.04a
• /
• pp.55-57
• /
• 2001

자바언어는 이질적인 네트워크 환경에서 프로그램 개발이 적합하도록 설계된 언어이다. 자바언어의 특징은 소프트웨어를 쉽게 개발하는데 유용한 것은 사실이지만, 성능상 제약이 따르게 된다. 즉 자바는 클래스 파일이 이동하여 JVM 환경에서 인터프리팅 되는 시스템이므로, 클래스 파일이 이동하며 실행되는 동안의 성능의 저하 없이 자바의 특징을 이용하려면 복잡한 최적화와 실행 시스템이 요구된다. 본 논문은 네트워크 상에서 동적으로 다운로드 되는 클래스 파일의 최적화에 있다. 클래스 파일이 인터프리팅 되는 시스템이 보다 적은 네트워크 로드를 가지고 실행할 수 있도록 하며, 효율적인 실행 속도를 보이도록 하는 것이다. 여기서는 Class Field Optimizer는 내부적으로 Bytecode Optimizer와 ClassGen을 이용하여 실행시간을 개선하고 전체 클래스 파일의 크기를 줄이게 된다. Bytecode Optimizer는 peephole 최적화를 수행하고, bytecode 의존적 최적화, 그리고 전역최적화를 행하게 된다. ClassGen은 클래스 파일의 포맷에 따라 bytecode를 분석하고 본래의 클래스 파일보다 작은 크기의 클래스 파일을 생성하게 된다. 최적화된 클래스 파일은 부분적으로 클래스 파일의 최적화를 가져와 전체 클래스 파일의 크기를 줄이고, 인터프리터를 통하여 실행될 때 수행 속도면에서 좀더 빠른 실행 속도를 가지게 된다.