• Journal of Information Processing Systems
• /
• v.5 no.4
• /
• pp.175-186
• /
• 2009

Power consumed by modern computer systems, particularly servers in data centers has almost reached an unacceptable level. However, their energy consumption is often not justifiable when their utilization is considered; that is, they tend to consume more energy than needed for their computing related jobs. Task scheduling in distributed computing systems (DCSs) can play a crucial role in increasing utilization; this will lead to the reduction in energy consumption. In this paper, we address the problem of scheduling precedence-constrained parallel applications in DCSs, and present two energy- conscious scheduling algorithms. Our scheduling algorithms adopt dynamic voltage and frequency scaling (DVFS) to minimize energy consumption. DVFS, as an efficient power management technology, has been increasingly integrated into many recent commodity processors. DVFS enables these processors to operate with different voltage supply levels at the expense of sacrificing clock frequencies. In the context of scheduling, this multiple voltage facility implies that there is a trade-off between the quality of schedules and energy consumption. Our algorithms effectively balance these two performance goals using a novel objective function and its variant, which take into account both goals; this claim is verified by the results obtained from our extensive comparative evaluation study.

• Wind and Structures
• /
• v.20 no.3
• /
• pp.363-388
• /
• 2015

For controlling the vibration of specific building structure with large span, a practical method for the design of MTMD was developed according to the characteristics of structures subjected to wind loads. Based on the model of analyzing wind-induced response of large-span structure with MTMD, the optimization method of multiple tuned mass dampers for large-span roof structures subjected to wind loads was established, in which the applicable requirements for strength and fatigue life of TMD spring were considered. According to the method, the controlled modes and placements of TMDs in MTMD were determined through the quantitative analysis on modal contribution to the wind-induced dynamic response of structure. To explore the characteristics of MTMD, the parametric analysis on the effects of mass ratio, damping ratio, central tuning frequency ratio and frequency range of MTMD, was performed in the study. Then the parameters of MTMD were optimized through genetic algorithm and the optimized MTMD showed good dynamic characteristics. The robustness of the optimized MTMD was also investigated.

• Wind and Structures
• /
• v.16 no.2
• /
• pp.157-178
• /
• 2013

Output-only modal parameter identification is based on the assumption that external forces on a linear structure are white noise. However, harmonic excitations are also often present in real structural vibrations. In particular, it has been realized that the use of forced acceleration responses without knowledge of external forces can pose a problem in the modal parameter identification, because an external force is imparted to its impulse acceleration response function. This paper provides a three-stage identification procedure as a solution to the problem of harmonic and white noise excitations in the acceleration responses of a linear dynamic system. This procedure combines the uses of the mode indicator function, the complex mode indication function, the enhanced frequency response function, an iterative rational fraction polynomial method and mode shape inspection for the correlation-related functions of the force-embedded acceleration responses. The procedure is verified via numerical simulation of a five-floor shear building and a two-dimensional frame and also applied to ambient vibration data of a large-span roof structure. Results show that the modal parameters of these dynamic systems can be satisfactorily identified under the requirement of wide separation between vibration modes and harmonic excitations.

• Coupled systems mechanics
• /
• v.7 no.3
• /
• pp.353-371
• /
• 2018

The Wavenumber or more accurately Wavenumber-FD approach was initially introduced for two-dimensional dynamic analysis of concrete gravity dam-reservoir systems. The technique was formulated in the context of pure finite element programming in frequency domain. Later on, a variation of the method was proposed which was referred to as Wavenumber-TD approach suitable for time domain type of analysis. Recently, it is also shown that Wavenumber-FD approach may be applied for three-dimensional dynamic analysis of concrete arch dam-reservoir systems. In the present study, application of its variation (i.e., Wavenumber-TD approach) is investigated for three-dimensional problems. The method is initially described. Subsequently, the response of idealized Morrow Point arch dam-reservoir system is obtained by this method and its special cases (i.e., two other well-known absorbing conditions) for time harmonic excitation in stream direction. All results for various considered cases are compared against the exact response for models with different values of normalized reservoir length and reservoir base/sidewalls absorptive conditions.

• Information and Communications Magazine
• /
• v.24 no.3
• /
• pp.76-86
• /
• 2007

This paper describes the RF performance issues of UE RF Transceiver for W-CDMA system based on 3GPP specifications. the parameters of transmitter and receiver are derived from the viewpoint of RF performance. In order for UE to achieve high performance, the transceiver performance requirements such as ACIR, EVM, Peak Code Domain Error, spectrum emission mask, frequency error stability and TX power control dynamic range for transmitter and reference sensitivity level, blocking characteristics, noise figure, ACS, linearity, AGC dynamic range for receiver are considered. On the basis of the required parameters, the UE RF transceiver is designed and then implemented. The evaluation of RF performance is accomplished through practical test scenarios.

• Advances in nano research
• /
• v.14 no.1
• /
• pp.67-80
• /
• 2023

Economic investigation is one of the main issues regarding the design and production of small-scale structures. This paper concerns the creation, implementation, and economic aspects of the cross-section profile of small-scale structures regarding the dynamic response of the free and forced vibration behavior of spinning nanoscale beams based on big data analysis. According to the financial analysis, the three practical non-uniform functions of cross-sections are compared to the uniform beam in the same weight and the equal material used. The previous studies reported that the uniform beams are more stable and contain a better frequency response based on the mechanical analysis. Still, concerning the economic investigation, which means the considered structures should have equal length and have the same weight in the aspect of material used, the conclusion can be different from the mechanical aspect. Consequently, in the current paper, the dynamic response along with computer technology as well as the big data analysis of the free and forced vibration of the nanobeam regarding the economic shape of the cross-section is scrutinized.

• Journal of Intelligence and Information Systems
• /
• v.5 no.1
• /
• pp.103-123
• /
• 1999

Input filtering as a preprocessing method is so much crucial to get good performance in time series forecasting. There are a few preprocessing methods (i.e. ARMA outputs as time domain filters, and Fourier transform or wavelet transform as time-frequency domain filters) for handling time series. Specially, the time-frequency domain filters describe the fractal structure of financial markets better than the time domain filters due to theoretically additional frequency information. Therefore, we, first of all, try to describe and analyze specially some issues on the effectiveness of different filtering methods from viewpoint of the performance of a neural network based forecasting. And then we discuss about neural network model architecture issues, for example, what type of neural network learning architecture is selected for our time series forecasting, and what input size should be applied to a model. In this study an input selection problem is limited to a size selection of the lagged input variables. To solve this problem, we simulate on analyzing and comparing a few neural networks having different model architecture and also use an embedding dimension measure as chaotic time series analysis or nonlinear dynamic analysis to reduce the dimensionality (i.e. the size of time delayed input variables) of the models. Throughout our study, experiments for integration methods of joint time-frequency analysis and neural network techniques are applied to a case study of daily Korean won / U. S dollar exchange returns and finally we suggest an integration framework for future research from our experimental results.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.45 no.1
• /
• pp.78-85
• /
• 2008

Most researches for power management have focused on increasing the utilization of system performance by scaling operating frequency or operating voltage. If operating frequency is changed frequently, it reduces the real system performance. To reduce power consumption, alternative approaches use the limited number of operating frequencies or set the smoothing frequencies during execution to increase the system performance, but they are not suitable for real time applications. To reduce power consumption and increase system performance for real time applications, this paper proposes a new power-aware schedule method by allocating operating frequencies and by setting smoothing frequencies. The algorithm predicts so that frequencies with continuous interval are mapped into discrete operating frequencies. The frequency smoothing reduces overheads of systems caused by changing operating frequencies frequently as well as power consumption caused by the frequency mismatch at a wide frequency interval. The simulation results show that the proposed algorithm reduces the power consumption up to 40% at maximum and 15% on average compared to the CC RT-DVS.

• Journal of Korea Multimedia Society
• /
• v.11 no.5
• /
• pp.641-650
• /
• 2008

As portable devices are used widely, power management algorithm is essential to extend battery use time on small-sized battery power. Although many methods have been proposed, they assumed the voltage transition overhead was negligible or was considered partially. However, the voltage transition overhead might not guarantee to schedule real-time tasks in portable multimedia systems. This paper proposes the adaptive power-aware algorithm to minimize the power consumption by considering the voltage transition overhead. It selects only a few discrete frequencies from the whole frequencies of a system and adjusts the interval between two consecutive frequencies based on the system utilization to reduce the number of frequency change. This algorithm saves the power consumption about 10 to 25 percent compared to a CC RT-DVS method and a frequency-smoothing method.

• Asia pacific journal of information systems
• /
• v.33 no.3
• /
• pp.737-767
• /
• 2023

With the availability of big customer data and advances in machine learning techniques, the prediction of customer behavior at the session-level has attracted considerable attention from marketing practitioners and scholars. This study aims to predict customer purchase conversion at the session-level by employing customer profile, transaction, and clickstream data. For this purpose, we develop a multimodal deep learning fusion model with dynamic and static features (i.e., DS-fusion). Specifically, we base page views within focal visist and recency, frequency, monetary value, and clumpiness (RFMC) for dynamic and static features, respectively, to comprehensively capture customer characteristics for buying behaviors. Our model with deep learning architectures combines these features for conversion prediction. We validate the proposed model using real-world e-commerce data. The experimental results reveal that our model outperforms unimodal classifiers with each feature and the classical machine learning models with dynamic and static features, including random forest and logistic regression. In this regard, this study sheds light on the promise of the machine learning approach with the complementary method for different modalities in predicting customer behaviors.