• Journal of KIISE:Computing Practices and Letters
• /
• v.16 no.6
• /
• pp.648-653
• /
• 2010

The non-metric multidimensional scaling (nMDS) is a method for analyzing the relation among objects by mapping them onto the Euclidean space. The nMDS is useful when it is difficult to use the concept of distance between pairs of objects due to non-metric dissimilarities between objects. The nMDS can be regarded as an optimization problem in which there are many local optima. Since the conventional nMDS algorithm utilizes the steepest descent method, it has a drawback in that the method can hardly find a better solution once it falls into a local optimum. To remedy this problem, in this paper, we applied the simulated annealing to the nMDS and proposed a new optimization algorithm which could search for a global optimum more effectively. We examined the algorithm using benchmarking problems and found that improvement rate of the proposed algorithm against the conventional algorithm ranged from 0.7% to 3.2%. In addition, the statistical hypothesis test also showed that the proposed algorithm outperformed the conventional one.

• KIPS Transactions on Software and Data Engineering
• /
• v.6 no.3
• /
• pp.155-160
• /
• 2017

Recently, Deep learning has been used successfully to solve many recognition problems. It has many advantages over existing machine learning methods that extract feature points through hand-crafting. Deep neural networks for human activity recognition split video data into frame images, and then classify activities by analysing the connectivity of frame images according to the time. But it is difficult to apply to actual problems which has small-scale activity classes. Because this situations has a problem of overfitting and insufficient training data. In this paper, we defined 5 type of small-scale human activities, and classified them. We construct video database using 700 video clips, and obtained a classifying accuracy of 74.00%.

• The Korean Journal of Applied Statistics
• /
• v.27 no.2
• /
• pp.249-261
• /
• 2014

Ever increasing "Big data" can only be effectively processed by parallel computing. Parallel computing refers to a high performance computational method that achieves effectiveness by dividing a big query into smaller subtasks and aggregating results from subtasks to provide an output. However, it is well-known that parallel computing does not achieve scalability which means that performance is improved linearly by adding more computers because it requires a very careful assignment of tasks to each node and collecting results in a timely manner. Hadoop is one of the most successful platforms to attain scalability. In this paper, we propose a measurement for Hadoop optimization by utilizing a Lorenz curve which is a proxy for the inequality of hardware resources. Our proposed index takes into account the intrinsic overhead of Hadoop systems such as CPU, disk I/O and network. Therefore, it also indicates that a given Hadoop can be improved explicitly and in what capacity. Our proposed method is illustrated with experimental data and substantiated by Monte Carlo simulations.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.39B no.10
• /
• pp.664-673
• /
• 2014

This paper analyzes the power allocation scheme to maximize the sum-rate for heterogeneous networks based on asynchronous time division duplex. We consider heterogeneous networks where a small cell exists in the macro cell coverage and the small cell and the macro cell share the same time-frequency resources. We formulate the optimization problem which maximizes the sum-rate of the heterogeneous network subject to the target signal-to-interference-plus-noise ratio. We analyze the feasible region in order for the optimal solution to exists and the optimal power allocation scheme for maximizing the sum-rate. Simulation results show that the proposed power allocation schemes outperform the maximum power transmission scheme.

• Smart Structures and Systems
• /
• v.15 no.5
• /
• pp.1373-1392
• /
• 2015

Placing sensors at appropriate locations is an important task in the design of an efficient structural health monitoring (SHM) system for a large-scale civil structure. In this paper, a hybrid optimization algorithm called virus monkey algorithm (VMA) based on the virus theory of evolution is proposed to seek the optimal placement of sensors. Firstly, the dual-structure coding method is adopted instead of binary coding method to code the solution. Then, the VMA is designed to incorporate two populations, a monkey population and a virus population, enabling the horizontal propagation between the monkey and virus individuals and the vertical inheritance of monkey's position information from the previous to following position. Correspondingly, the monkey population in this paper is divided into the superior and inferior monkey populations, and the virus population is divided into the serious and slight virus populations. The serious virus is used to infect the inferior monkey to make it escape from the local optima, while the slight virus is adopted to infect the superior monkey to let it find a better result in the nearby area. This kind of novel virus infection operator enables the coevolution of monkey and virus populations. Finally, the effectiveness of the proposed VMA is demonstrated by designing the sensor network of the Canton Tower, the tallest TV Tower in China. Results show that innovations in the VMA proposed in this paper can improve the convergence of algorithm compared with the original monkey algorithm (MA).

• Journal of the Korea Society for Simulation
• /
• v.17 no.1
• /
• pp.17-22
• /
• 2008

Double-binary turbo code has been adopted as an error control code of various future communication systems including wireless metropolitan area networks(WMAN) due to its powerful error correction capability. One of the components affecting the performance of turbo code is internal interleaver. In 802.16 d/e system, an almost regular permutation(ARP) interleaver has been included as a part of specification, however it seems that the interleaver is not optimized in terms of decoding performance. In this paper, we propose three optimization methods for the interleaver based on spatial distance, spread and minimum distance between original and interleaved sequence. We find optimized interleaving parameters for each optimization method and evaluate the performances of the proposed methods by computer simulation under additive white Gaussian noise(AWGN) channel. Optimized parameters can provide up to 1.0 dB power gain over the conventional method and furthermore the obtainable gain does not require any additional hardware complexity.

• Geomechanics and Engineering
• /
• v.20 no.3
• /
• pp.191-205
• /
• 2020

Soil shear strength parameters play a remarkable role in designing geotechnical structures such as retaining wall and dam. This study puts an effort to propose two accurate and practical predictive models of soil shear strength parameters via hybrid artificial neural network (ANN)-based models namely genetic algorithm (GA)-ANN and particle swarm optimization (PSO)-ANN. To reach the aim of this study, a series of consolidated undrained Triaxial tests were conducted to survey inherent strength increase due to addition of polypropylene fibers to sandy soil. Fiber material with different lengths and percentages were considered to be mixed with sandy soil to evaluate cohesion (as one of shear strength parameter) values. The obtained results from laboratory tests showed that fiber percentage, fiber length, deviator stress and pore water pressure have a significant impact on cohesion values and due to that, these parameters were selected as model inputs. Many GA-ANN and PSO-ANN models were constructed based on the most effective parameters of these models. Based on the simulation results and the computed indices' values, it is observed that the developed GA-ANN model with training and testing coefficient of determination values of 0.957 and 0.950, respectively, performs better than the proposed PSO-ANN model giving coefficient of determination values of 0.938 and 0.943 for training and testing sets, respectively. Therefore, GA-ANN can provide a new applicable model to effectively predict cohesion of fiber-reinforced sandy soil.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.37 no.5A
• /
• pp.325-338
• /
• 2012

In this paper, a radio resource allocation scheme for a multi-hop relay transmission in cognitive radio (CR) system is proposed to support the employment of relay nodes in IEEE 802.22 standard for wireless regional area network (WRAN). An optimization problem is formulated to maximize the number of serving secondary users (SUs) under system constraints such as time-divided frame structure for multiplexing and a single resource-unit to every relay-hop. However, due to mathematical complexity, the optimization problem is solved with a sub-optimal manner instead, which takes three steps in the order of user selection, relay/path selection, and frequency selection. In the numerical analysis, this proposed solution is evaluated in terms of service rate denoting as the ratio of the number of serving SUs to the number of service-requesting SUs. Simulation results show the condition of adopting multi-hop relay and the optimum number of relaying hops by comparing with the performance of 1-hop system.

• The Transactions of the Korea Information Processing Society
• /
• v.6 no.5
• /
• pp.1393-1402
• /
• 1999

This paper proposes an efficient principal feature extraction of the image data using neural networks of a new learning algorithm. The proposed learning algorithm is a backpropagation(BP) algorithm based on the steepest descent and dynamic tunneling. The BP algorithm based on the steepest descent is applied for high-speed optimization, and the BP algorithm based on the dynamic tunneling is also applied for global optimization. Converging to the local minimum by the BP algorithm of steepest descent, the new initial weights for escaping the local minimum is estimated by the BP algorithm of dynamic tunneling. The proposed algorithm has been applied to the 3 image data of 12${\times}$12pixels and the Lenna image of 128${\times}$128 pixels respectively. The simulation results shows that the proposed algorithm has better performances of the convergence and the feature extraction, in comparison with those using the Sanger method and the Foldiak method for single-layer neural networks and the BP algorithm for multilayer neural network.

• Journal of the Korean Institute of Illuminating and Electrical Installation Engineers
• /
• v.23 no.12
• /
• pp.154-161
• /
• 2009

A fuel cell power plant is a very complex system which has various control loops with some non-linearity. For control of a fuel cell power plant, dynamic models of fuel cell stacks have been developed and simplified process flow diagrams of a fuel cell power plant has been presented. Using such a model of a Molten Carbonate Fuel Cell (MCFC) power plant, this paper deals with development of an intelligent setpoint reference governor (I-SRG) to find the optimal setpoints and feed forward control inputs for the plant power demand. The I-SRG is implemented with neural network by using Particle Swarm Optimization (PSO) algorithm based on system constraints and performance objectives. The feasibility of the I-SRG is shown through simulation of an MCFC power plant for tracking control of its power demand.