• 산업공학
• /
• 제12권2호
• /
• pp.312-318
• /
• 1999

This paper considers the cell loss probability(CLP) in a multiplexer with the superposition of heterogeneous ON/OFF sources. The input traffic is composed of k classes. Traffic of class i is the superposition of M_(i) ON/OFF sources. Recently, the method based on the Markov modulated deterministic process(MMDP) is presented. Basically, it is the discretized model of stochastic fluid flow process(SFFP) and gives the CLP very fast, but under-estimates the CLP especially when the value of estimated CLP is very low. This paper develops the discretized model of Markov modulated Poisson process(MMPP). It is a special type of switched batch Bernoulli process(SBBP). Combining the transition probability matrix of MMDP and SBBP according to the state which is characterized by the arrival rate, this paper presents hybrid algorithm. The hybrid algorithm gives better estimate of CLP than that of MMDP and faster than SBBP.

• 한국산업정보학회논문지
• /
• 제2권2호
• /
• pp.219-228
• /
• 1997

The Fano metric is the maximum likelihood decoding choice for convlutional code for binary symmetric channel. The Fano metric assumes that it has previous knowledge of channel error probability. However, the bit errors in real channel occur in bursts and the channel error probability can not be known exactly. Thus, the Fano metric is not the maximum likelihood choice for bursty-noise channel. In this paper universal metri which dose not require the previous knowlege of the channel transition probability is used for sequential decoding. It is shown that the complexity of the universal is much less than that of the Fano metric bursty-noise channel, since it is estimated on a branch by branch basis.

• International Journal of Naval Architecture and Ocean Engineering
• /
• 제2권3호
• /
• pp.119-126
• /
• 2010

The response of ship roll oscillation under random ice impulsive loads modeled by Poisson arrival process is very important in studying the safety of ships navigation in cold regions. Under both external and parametric random excitations the evolution of the probability density function of roll motion is evaluated using the path integral (PI) approach. The PI method relies on the Chapman-Kolmogorov equation, which governs the response transition probability density functions at two close intervals of time. Once the response probability density function at an early close time is specified, its value at later close time can be evaluated. The PI method is first demonstrated via simple dynamical models and then applied for ship roll dynamics under random impulsive white noise excitation.

• 대한임베디드공학회논문지
• /
• 제10권5호
• /
• pp.265-272
• /
• 2015

This paper presents an improved feature extraction method to achieve a good performance in the detection of splicing forged images. Strong edges caused by the image splicing destroy the statistical dependencies between parent and child subbands in the wavelet domain. We analyze the co-occurrence probability matrix of parent and child subbands in the wavelet domain, and calculate the statistical moments from two-dimensional characteristic function of the co-occurrence matrix. The extracted features are used as the input of SVM classifier. Experimental results show that the proposed method obtains a good performance with a small number of features compared to the existing methods.

• 한국철도학회:학술대회논문집
• /
• 한국철도학회 2009년도 춘계학술대회 논문집
• /
• pp.1-5
• /
• 2009

Recently, as a method of green-development and reduction of carbon dioxide emission, increased interest has been focused on a railway. Furthermore, an intensive study has been processed on capabilities of maintenance activities, economic efficiency of maintenance on rail structure and a design of railway structure as well as the development of materials. The purpose of this paper is to develop a deteriorated model of PC Beam Bridge due to timely changes and maintenance activities. Typically, there is definite difference between maintained bridges and non-maintained bridges. As a result of proper maintenance activity, a life time of a structure can be enhanced. In this study, we will research and analyze structures with ongoing maintenance. We will also process same procedures on structures without maintenance. Therefore, we can establish the significant role in a conditional change of a structure. Based on a study, we accomplish the development of a condition-deteriorated model. To develop deteriorated model of PC Beam Bridge, We apply Marcov Theory and develop a transition probability to show the life time of bridge. This study will provide a great benefit to decision making for maintenance activities on the railway bridges for future.

• 대한산업공학회지
• /
• 제38권3호
• /
• pp.208-219
• /
• 2012

As the number of end-of-life vehicles (ELVs) has kept increasing, the management of ELV has also become one of the academic research focuses and European Union recently adopted the directive on ELVs. For the stakeholders has become a principle agent of dealing with all about ELVs, it is relevant investment decision to set up and to decide high-cost ELVs entity locations and to forecast future ELVs' amount in advance. In this paper, transition probability matrixes between months are made by using Markov Chain and the number of ELVs is predicted with them. This study will perform a great role as a fundamental material in Korea where just started having interests about recycling resources and studies related to the topic. Moreover, the forecasting method developed for this research can be adopted for other enhancements in different but comparable situations.

• 한국지하수토양환경학회지:지하수토양환경
• /
• 제21권3호
• /
• pp.21-34
• /
• 2016

In the present study, we applied various machine learning techniques comparatively for prediction of subsurface structures based on multiple secondary information (i.e., well-logging data). The machine learning techniques employed in this study are Naive Bayes classification (NB), artificial neural network (ANN), support vector machine (SVM) and logistic regression classification (LR). As an alternative model, conventional hidden Markov model (HMM) and modified hidden Markov model (mHMM) are used where additional information of transition probability between primary properties is incorporated in the predictions. In the comparisons, 16 boreholes consisted with four different materials are synthesized, which show directional non-stationarity in upward and downward directions. Futhermore, two types of the secondary information that is statistically related to each material are generated. From the comparative analysis with various case studies, the accuracies of the techniques become degenerated with inclusion of additive errors and small amount of the training data. For HMM predictions, the conventional HMM shows the similar accuracies with the models that does not relies on transition probability. However, the mHMM consistently shows the highest prediction accuracy among the test cases, which can be attributed to the consideration of geological nature in the training of the model.

• 한국멀티미디어학회논문지
• /
• 제12권12호
• /
• pp.1819-1826
• /
• 2009

To enhance the automatic performance of existing predicting and planning algorithms that require a predefined probability of the states' transition, this paper proposes a multiple sequence generation system. When interacting with unknown environments, a virtual agent needs to decide which action or action order can result in a good state and determine the transition probability based on the current state and the action taken. We describe a sequential behavior generation method motivated from the change in the agent's state in order to help the virtual agent learn how to adapt to unknown environments. In a sequence learning process, the sensed states are grouped by a set of proposed motivation filters in order to reduce the learning computation of the large state space. In order to accomplish a goal with a high payoff, the learning agent makes a decision based on the observation of states' transitions. The proposed multiple sequence behaviors generation system increases the complexity and heightens the automatic planning of the virtual agent for interacting with the dynamic unknown environment. This model was tested in a virtual library to elucidate the process of the system.

• 한국음향학회지
• /
• 제10권2호
• /
• pp.29-35
• /
• 1991

본 논문에서는 변이음 단위의 Hidden Markov Model (HMM)을 이용하여 고립단어를 인식하는 방법을 논한다. 변이음 단위로 HMM을 구성하여 변이음 사전을 만들고, 이 변이음 사전을 이용하여 단어 사전을 구성한다. 변이음 HMM을 이용하여 단어를 구성하려면 변이음 간의 천이확률이 계산되어야 하므로 본 연구에서는 변이음 간의 천이 확률의 영향을 측정하여 그 변이음으로 이루어지는 임의의 단어를 적응없이 적은 수의 적응 데이터로 단어모델을 구성 인식하는 것을 설명한다. 비교를 위하여 단어인식 HMM으로 인식 실험을 한 결과, 변이음 단위 HMM이 적은 기억 용량과 적은 데이터의 훈련으로 단어단위 HMM 이상의 인식률을 얻을 수 있음을 보였다.

• 대한기계학회:학술대회논문집
• /
• 대한기계학회 2008년도 추계학술대회A
• /
• pp.1788-1793
• /
• 2008

Classical molecular dynamics simulations (MDS) were conducted to simulate nano-sized cluster collisions with a weakly attractive static surface. Energy exchanges associated with the cluster collision and the adhesion probability are discussed. Routes of the energy exchanges and the kinetic energy loss are vastly altered in their mode according to the cluster incident velocity. In the elastic collision regime ($V_0$<0.1), most incident kinetic energy is recovered into the rebounding kinetic energy, but a little loss in the incident kinetic energy causes the cluster adhesion. Dissipated kinetic energy is converted into the rotational energy. In the weakly plastic collision regime (0.1<$V_0$<0.3), the transition from elastic to plastic collision occurs, and a large part of the released potential energy is converted into rebounding translational energy. For strongly plastic collisions ($V_0$>0.3), permanent cluster deformation occurs with extensive collapse of the lattice structure inducing a solid-to-solid phase transition; moreover, most of the cluster kinetic energy is converted into cluster potential and thermal energy.