• Journal of applied mathematics & informatics
• /
• v.5 no.2
• /
• pp.403-426
• /
• 1998

We present an algorithm to find an approximation for the stationary distribution for the general ergodic spatially-inhomogeneous block-partitioned upper Hessenberg form. Our approximation makes use of an associated upper block-Hessenberg matrix which is spa-tially homogeneous except for a finite number of blocks. We treat the MAP/G/1 retrial queue and the retrial queue with two types of customer as specific instances and give some numerical examples. The numerical results suggest that our method is superior to the ordinary finite-truncation method.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.39 no.2
• /
• pp.1-10
• /
• 2016

In this paper, we present a new way to derive the mean cycle time of the G/G/m failure prone queue when the loading of the system approaches to zero. The loading is the relative ratio of the arrival rate to the service rate multiplied by the number of servers. The system with low loading means the busy fraction of the system is low. The queueing system with low loading can be found in the semiconductor manufacturing process. Cluster tools in semiconductor manufacturing need a setup whenever the types of two successive lots are different. To setup a cluster tool, all wafers of preceding lot should be removed. Then, the waiting time of the next lot is zero excluding the setup time. This kind of situation can be regarded as the system with low loading. By employing absorbing Markov chain model and renewal theory, we propose a new way to derive the exact mean cycle time. In addition, using the proposed method, we present the cycle times of other types of queueing systems. For a queueing model with phase type service time distribution, we can obtain a two dimensional Markov chain model, which leads us to calculate the exact cycle time. The results also can be applied to a queueing model with batch arrivals. Our results can be employed to test the accuracy of existing or newly developed approximation methods. Furthermore, we provide intuitive interpretations to the results regarding the expected waiting time. The intuitive interpretations can be used to understand logically the characteristics of systems with low loading.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1995.09a
• /
• pp.99-107
• /
• 1995

The Phase-type(PH) distribution, defined as a distribution of the time until the absorption in a finite continuous-time Markov chain state with one absorbing state, has been widely used for various stochastic modelling. But great computational burdens often make us hesitate to apply PH methods. In this paper, we propose a seemingly efficient approximation method for phase type distributions. We first describe methods to bound the first passage time distribution in continuous-time Markov chains. Next, we adapt these bounding methods to approximate phase-tupe distributions. Numerical computation results are given to verify their efficiency.

• Communications for Statistical Applications and Methods
• /
• v.29 no.5
• /
• pp.513-532
• /
• 2022

In this paper, the use of Bayes factors and the deviance information criterion for model selection are compared in a Bayesian accelerated life testing setup. In Bayesian accelerated life testing, the most used tool for model comparison is the deviance information criterion. An alternative and more formal approach is to use Bayes factors to compare models. However, Bayesian accelerated life testing models with more than one stressor often have mathematically intractable posterior distributions and Markov chain Monte Carlo methods are employed to obtain posterior samples to base inference on. The computation of the marginal likelihood is challenging when working with such complex models. In this paper, methods for approximating the marginal likelihood and the application thereof in the accelerated life testing paradigm are explored for dual-stress models. A simulation study is also included, where Bayes factors using the different approximation methods and the deviance information are compared.

• Transactions of the Korean Society of Machine Tool Engineers
• /
• v.13 no.2
• /
• pp.81-87
• /
• 2004

This paper deals with a new approach to tolerance optimization problems. Optimal tolerance allotment problems can be formulated as stochastic optimization problems. Most schemes to solve the stochastic optimization problems have been found to exhibit difficulties in multivariate integration of the probability density function. As a typical example of stochastic optimization the optimal tolerance allotment problem has the same difficulties. In this stochastic model, manufacturing system is represented by Gauss-Markov stochastic process and the manufacturing unit availability is characterized for realistic optimization modeling. The new algorithm performed robustly for a large deviation approximation. A significant reduction in computation time was observed compared to the results obtained in previous studies.

• Journal of the Korean Statistical Society
• /
• v.36 no.2
• /
• pp.321-333
• /
• 2007

In the proportional hazard model with the beta process prior, the posterior computation with the discrete approximation is considered. The time period of interest is partitioned by small intervals. On each partitioning interval, the likelihood is approximated by that of a binomial experiment and the beta process prior is by a beta distribution. Consequently, the posterior is approximated by that of many independent binomial model with beta priors. The analysis of the leukemia remission data is given as an example. It is illustrated that the length of the partitioning interval affects the posterior and one needs to be careful in choosing it.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.15 no.25
• /
• pp.75-82
• /
• 1992

Standard acceptance sampling plans models the production pricess as a sequence of independent identically distributed Beruoulli random variables. However, the quality of items sampled sequentially from an ongoing production process of ten exhibits statistical dependency that is not accounted for in standard acceptance sampling plans. In this paper, a dependent production process is modelled as an ARMA process and as a two-state Markov chain. A simulation study of each is performed. A comparison of the probability of acceptance is done for the simulation method and for the approximation method.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.7 no.1
• /
• pp.81-98
• /
• 2013

Since Probabilistic Latent Semantic Analysis (PLSA) and Latent Dirichlet Allocation (LDA) were introduced, many revised or extended topic models have appeared. Due to the intractable likelihood of these models, training any topic model requires to use some approximation algorithm such as variational approximation, Laplace approximation, or Markov chain Monte Carlo (MCMC). Although these approximation algorithms perform well, training a topic model is still computationally expensive given the large amount of data it requires. In this paper, we propose a new method, called non-simultaneous sampling deactivation, for efficient approximation of parameters in a topic model. While each random variable is normally sampled or obtained by a single predefined burn-in period in the traditional approximation algorithms, our new method is based on the observation that the random variable nodes in one topic model have all different periods of convergence. During the iterative approximation process, the proposed method allows each random variable node to be terminated or deactivated when it is converged. Therefore, compared to the traditional approximation ways in which usually every node is deactivated concurrently, the proposed method achieves the inference efficiency in terms of time and memory. We do not propose a new approximation algorithm, but a new process applicable to the existing approximation algorithms. Through experiments, we show the time and memory efficiency of the method, and discuss about the tradeoff between the efficiency of the approximation process and the parameter consistency.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.4 no.3
• /
• pp.549-555
• /
• 2000

We propose an approximate algorithm to analyze the queuing system with n bursty and heterogeneous arrival processes. Each input process is modeled by Interrupted Bernoulli Process(IBP). We approximate N arrival processes by a single state variable and subsequently simplify the transition probability matrix of the Markov chain associated with these N arrival processes. Using this single state variable of arrival processes, we describe the state of the queuing system and analyze the system numerically with the reduced transition probability matrix. We compute the queue length distribution, the delay distribution, and the loss probability. Comparisons with simulation data show that the approximation algorithm has a good accuracy.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.33 no.10
• /
• pp.1163-1170
• /
• 2009

Reliability analysis is of great importance in the advanced product design, which is to evaluate reliability due to the associated uncertainties. There are three types of uncertainties: the first is the aleatory uncertainty which is related with inherent physical randomness that is completely described by a suitable probability model. The second is the epistemic uncertainty, which results from the lack of knowledge due to the insufficient data. These two uncertainties are encountered in the input variables such as dimensional tolerances, material properties and loading conditions. The third is the metamodel uncertainty which arises from the approximation of the response function. In this study, an integrated method for the reliability analysis is proposed that can address all these uncertainties in a single Bayesian framework. Markov Chain Monte Carlo (MCMC) method is employed to facilitate the simulation of the posterior distribution. Mathematical and engineering examples are used to demonstrate the proposed method.