• 대한전기학회논문지:전력기술부문A
• /
• 제53권10호
• /
• pp.529-535
• /
• 2004

Load forecasting is essential in the electricity market for the participants to manage the market efficiently and stably. A wide variety of techniques/algorithms for load forecasting has been reported in many literatures. These techniques are as follows: multiple linear regression, stochastic time series, general exponential smoothing, state space and Kalman filter, knowledge-based expert system approach (fuzzy method and artificial neural network). These techniques have improved the accuracy of the load forecasting. In recent 10 years, many researchers have focused on artificial neural network and fuzzy method for the load forecasting. In this paper, we propose an algorithm of a hybrid load forecasting method using fuzzy linear regression and general exponential smoothing and considering the sensitivities of the temperature. In order to consider the lower load of weekends and Monday than weekdays, fuzzy linear regression method is proposed. The temperature sensitivity is used to improve the accuracy of the load forecasting through the relation of the daily load and temperature. And the normal load of weekdays is easily forecasted by general exponential smoothing method. Test results show that the proposed algorithm improves the accuracy of the load forecasting in 1996.

• 융합보안논문지
• /
• 제8권1호
• /
• pp.91-100
• /
• 2008

The previous works in sensor networks security have focused on the aspect of confidentiality, authentication and integrity based on cryptographic primitives. There has been no prior work to assess the survivability in systematic way. Accordingly, this paper presents a survivability model of wireless sensor networks using software rejuvenation for dual adaptive cluster head. The survivability model has state transition to reflect status of real wireless sensor networks. In this paper, we only focus on a survivability model which is capable of describing cluster head compromise in the networks and able to switch over the redundant cluster head in order to increase the survivability of that cluster. Second, this paper presents how to enhance the survivability of sensor networks using software rejuvenation methodology for dual cluster head in wireless sensor network. We model and analyze each cluster as a stochastic process based on Semi Markov Process (SMP) and Discrete Time Markov Chain (DTMC). The proof of example scenarios and numerical analysis shows the feasibility of our approach.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제14권8호
• /
• pp.3519-3533
• /
• 2020

Tuberculosis is a chronic and delayed infection which is easily experienced by young people. According to the statistics of the World Health Organization (WHO), there are nearly ten million fell ill with tuberculosis and a total of 1.5 million people died from tuberculosis in 2018 (including 251000 people with HIV). Tuberculosis is the largest single infectious pathogen that leads to death. In order to help doctors with tuberculosis diagnosis, we compare the tuberculosis classification abilities of six popular convolutional neural network (CNN) models in the same data set to find the best model. Before training, we optimize three parts of CNN to achieve better results. We employ sigmoid function to replace the step function as the activation function. What's more, we use binary cross entropy function as the cost function to replace traditional quadratic cost function. Finally, we choose stochastic gradient descent (SGD) as gradient descent algorithm. From the results of our experiments, we find that Densenet121 is most suitable for tuberculosis diagnosis and achieve a highest accuracy of 0.835. The optimization and expansion depend on the increase of data set and the improvements of Densenet121.

• 제어로봇시스템학회논문지
• /
• 제18권1호
• /
• pp.35-41
• /
• 2012

This paper proposes a moving object detection algorithm for active camera system that can be applied to mobile robot and intelligent surveillance system. Most of moving object detection algorithms based on a stationary camera system. These algorithms used fixed surveillance system that does not consider the motion of the background or robot tracking system that track pre-learned object. Unlike the stationary camera system, the active camera system has a problem that is difficult to extract the moving object due to the error occurred by the movement of camera. In order to overcome this problem, the motion of the camera was compensated by using SURF and Pseudo Perspective model, and then the moving object is extracted efficiently using stochastic Label Cluster transport model. This method is possible to detect moving object because that minimizes effect of the background movement. Our approach proves robust and effective in terms of moving object detection in active camera system.

• 전기학회논문지
• /
• 제61권6호
• /
• pp.783-790
• /
• 2012

Electric Vehicles(EVs) and Plug-in Hybrid Electric Vehicles(PHEVs) which have the grid connection capability, represent an important power system issue of charging demands. Analyzing impacts EVs charging demands of the power system such as increased peak demands, developed by means of modeling a stochastic distribution of charging and a demand dispatch calculation. Optimization processes proposed to determine optimal demand distribution portions so that charging costs and demand can possibly be managed. In order to solve the problems due to increasing charging demand at the peak time, alternative electricity rate such as Time-of-Use(TOU) rate has been in effect since last year. The TOU rate would in practice change the tendencies of charging time at the peak time. Nevertheless, since it focus only minimizing costs of charging from owners of the EVs, loads would be concentrated at times which have a lowest charging rate and would form a new peak load. The purpose of this paper is that to suggest a scenario of load leveling for a power system operator side. In case study results, the vehicles as regular load with time constraints, battery charging patterns and changed daily demand in the charging areas are investigated and optimization results are analyzed regarding cost and operation aspects by determining optimal demand distribution portions.

• Industrial Engineering and Management Systems
• /
• 제12권3호
• /
• pp.207-223
• /
• 2013

Revenue management (RM) has been widely used to model products characterized as perishable. Classical RM model assumed that price is the sole factor in the model. Thus price adjustment becomes a crucial and costly factor in business. In this paper, an optimal pricing model is developed based on minimization of soft customer cost, one kind of price adjustment cost and is solved by Lagrange multiplier method. It is formed by expected discounted revenue/bid price integrating quantity-based RM and pricing-based RM. Quantity-based RM consists of two capacity models, namely, booking limit and overbooking. Booking limit, built by assuming uncertain customer arrival, decides the optimal capacity allocation for two market segments. Overbooking determines the level of accepted order exceeding capacity to anticipate probability of cancellation. Furthermore, pricing-based RM models occupancy/demand rate influenced by internal and competitor price changes. In this paper, a mathematical model based on game theoretic approach is developed for two conditions of deterministic and stochastic demand. Based on the equilibrium point, the best strategy for both hotels can be determined.

• 대한산업공학회지
• /
• 제33권2호
• /
• pp.191-200
• /
• 2007

We propose the QoS (Quality of Service) assurance model for AMR (Adaptive MultiRate) voice users considering the capacity and service quality jointly in downlink WCDMA system. For this purpose, we introduce a new system performance measure and the number-based AMR mode allocation scheme. The proposed number-based AMR mode allocation can be operated only with the information of total number of ongoing users. Therefore, it can be more simply implemented than the existing power-based allocation. The proposed system performance measure considers the stochastic variations of AMR modes of ongoing users and can be analytically obtained using CTMC (Continuous Time Markov Chain) modeling. In order to validate the proposed analytical model, a discrete event-based simulation model is also developed. The performance measure obtained from the analytical model is in agreement with the simulation results and is expected to be useful for parameter optimization.

• Advances in concrete construction
• /
• 제9권5호
• /
• pp.469-478
• /
• 2020

This study aims to establish a new methodological framework for the evaluation of the evolution of the reliability of plain concrete for pavement vs number of cycles under flexural fatigue loading. According to the framework, a new method calculating the reliability was proposed through probability simulation in order to describe a random accumulation of fatigue damage, which combines reliability theory, one-to-one probability density functions transformation technique, cumulative fatigue damage theory and Weibull distribution theory. Then the statistical analysis of flexural fatigue performance of cement concrete tested was carried out utilizing Weibull distribution. Ultimately, the reliability for the tested cement concrete was obtained by the proposed method. Results indicate that the stochastic evolution behavior of concrete materials under fatigue loading can be captured by the established framework. The flexural fatigue life data of concrete at different stress levels is well described utilizing the two-parameter Weibull distribution. The evolution of reliability for concrete materials tested in this study develops by three stages and may corresponds to develop stages of cracking. The proposed method may also be available for the analysis of degradation behaviors under non-fatigue conditions.

• Steel and Composite Structures
• /
• 제37권1호
• /
• pp.1-14
• /
• 2020

Composite beams with corrugated steel webs represent a new innovative system which has emerged in the past decade for medium span in the construction technology. The use of composite beams with corrugated steel webs results in a range of benefits, including flexible spaces and reduced foundation costs in the construction technology. The thin corrugated web affords a significant weight reduction of these beams, compared with hot-rolled or welded ones. In the current research, an optimal designed I-girder beam with corrugated web has been proposed to improve the structural performance of continuous composite girder under bending moment. The experimental program has been conducted for six simply supported composite beams with different loading conditions. The tested specimens are designed by using one of the stochastic techniques called hunting search algorithm. In the optimization process, besides the thickness of concrete slab and studs, corrugated web properties are considered as design variables. The design constraints are respectively implemented from Eurocode 3, BS-8110 and DIN 18-800 Teil-1. The last part of the study focuses on performing a numerical study on composite beams by utilizing finite element analysis and the bending behavior of steel girders with corrugated webs experimentally and numerically verified the results. A nonlinear analysis was carried out using the finite element software ANSYS on the composite beams which were modelled using the elements ten-node high order quadrilateral type.

• 한국정보시스템학회지:정보시스템연구
• /
• 제12권2호
• /
• pp.1-20
• /
• 2003

This paper deals with the application of the genetic algorithm to the technical trading rule of the stock market. MACD(Moving Average Convergence & Divergence) and the Stochastic techniques are widely used technical trading rules in the financial markets. But, it is necessary to determine the parameters of these trading rules in order to use the trading rules. We use the genetic algorithm to obtain the appropriate values of the parameters. We use the daily KOSPI data of eight years during January 1995 and October 2002 as the experimental data. We divide the total experimental period into learning period and testing period. The genetic algorithm determines the values of parameters for the trading rules during the teaming period and we test the performance of the algorithm during the testing period with the determined parameters. Also, we compare the return of the genetic algorithm with the returns of buy-hold strategy and risk-free asset. From the experiment, we can see that the genetic algorithm outperforms the other strategies. Thus, we can conclude that genetic algorithm can be used successfully to the technical trading rule.