Hidden semi-Markov model-based methodology for multi-sensor equipment health diagnosis and prognosis

This paper presents an integrated platform for multi-sensor equipment diagnosis and prognosis. This integrated framework is based on hidden semi-Markov model (HSMM). Unlike a state in a standard hidden Markov model (HMM), a state in an HSMM generates a segment of observations, as opposed to a single observation in the HMM. Therefore, HSMM structure has a temporal component compared to HMM. In this framework, states of HSMMs are used to represent the health status of a component. The duration of a health state is modeled by an explicit Gaussian probability function. The model parameters (i.e., initial state distribution, state transition probability matrix, observation probability matrix, and health-state duration probability distribution) are estimated through a modified forward–backward training algorithm. The re-estimation formulae for model parameters are derived. The trained HSMMs can be used to diagnose the health status of a component. Through parameter estimation of the health-state duration probability distribution and the proposed backward recursive equations, one can predict the useful remaining life of the component. To determine the “value” of each sensor information, discriminant function analysis is employed to adjust the weight or importance assigned to a sensor. Therefore, sensor fusion becomes possible in this HSMM based framework.     

The Australian National Bushfire model project

The National Bushfire Research Unit has developed a personal computer based model to act as a shell for a decision support system for bushfire management. To date the fire spread module has been developed, and there is a continuing programme to incorporate expert systems. The model is designed for real-time operation. Inputs to the model combine a geographic information system (GIS) with real-time data acquisition and data assimilation through modem connections to telephone lines. Outputs are graphical and allow the fire-front position, which is generated using Huygens' principle, to be examined at any desired scale.     

基于二维真实隐马氏模型的手写体文字识别算法

近几年来，人们采用各种方法试图将1D隐马氏模型(HMM)^[2]推广到2D隐马氏模型。令人失望的是由于在建立合适的2D模型及其计算上的复杂度问题上存在困难，前面的尝试都没有得到一个真实的2DHMM．本文对于应用真实2D隐马氏模型(隐马氏网格随机场HMMRF)^[1,4]进行手写字符识别问题提出新的框架，针对文献[1]中的单点最优算法给出局部最优的译码算法。HMMRF模型是1D隐马氏模型到2D的扩展，能更好的描述字符的2D特性。HMMRF在字符识别中的应用具有两个相——学习相和译码相。在学习相和译码相中我们的最优标准是基于极大边缘后验概率的。不过，在涉及到2D模型中的计算问题时，对模型做出某些简单化的假设是必要的。本文用到的方法对于在合理的模型假设下解决手写字符识别问题呈现了很大的潜力。     

Statistical inference for the nonparametric and semiparametric hidden Markov model via the composite likelihood approach

In this paper, we propose a new estimation method for a nonparametric hidden Markov model(HMM), in which both the emission model and the transition matrix are nonparametric, and a semiparametric HMM, in which the transition matrix is parametric while emission models are nonparametric. The estimation is based on a novel composite likelihood method, where the pairs of consecutive observations are treated as independent bivariate random variables. Therefore, the model is transformed into a mixture ...     

A prognosis model for wear prediction based on oil-based monitoring

This paper reports on the development of a wear prediction model based on stochastic filtering and hidden Markov theory. It is assumed that observations at discrete time points are available such as metal concentrations from oil-based monitoring, which are related to the true underlying state of the system which is unobservable. The system state is represented by a generic term of wear which is modelled by a continuous hidden Markov Chain using a Beta distribution. We formulated a recursive model to predict the current and future system state given past observed monitoring information to date. The model is useful to wear-based monitoring such as oil analysis. Numerical examples are presented in the paper based on simulated and real data.     

An extended stochastic failure model for a system subject to random shocks

In this article, a stochastic failure model for a system subject to a random shock process is studied. It is assumed that a fatal shock results in an immediate system failure, whereas a non-fatal shock may increase the susceptibility of the system to failure. The lifetime distribution of the system and its failure rate function are derived, and the effect of environmental factors on the failure process of the system is also investigated. Lifetimes of systems operated under different environmental conditions are stochastically compared.     

Behavior prediction of washing system in a paper industry using GA and fuzzy lambda–tau technique

Availability analysis has been an important issue in the design field of any Industrial system as the system structure has become more complicated. Also, the system availability is affected by many factors such as design, manufacturing, installation, etc., and so it may be extremely difficult to model, analyze and predict the failure behavior of the system. The purpose of this paper is to develop a new approach for computing various performance measures, namely reliability, availability, MTBF (mean time between failures), ENOF (expected number of failures), failure rate and repair time, for any industrial system. In the proposed approach, the failure rates and repair times of all constituent components are obtained using genetic algorithms and then various performance measures are computed using fuzzy lambda–tau methodology. Washing system, the major part of paper industry is the subject of study. The interactions among the working components are modeled using Petri nets. Failure and repair rates are represented using triangular fuzzy numbers as they allow expert opinion, linguistic variables, operating conditions, uncertainty and imprecision in reliability information to be incorporated into system model. Based on calculated reliability parameters, a structured framework has been developed that may help the maintenance engineers to analyze and predict the system behavior.     

Belief rule-based system for portfolio optimisation with nonlinear cash-flows and constraints

A belief rule-based (BRB) system is a generic nonlinear modelling and inference scheme. It is based on the concept of belief structures and evidential reasoning (ER), and has been shown to be capable of capturing complicated nonlinear causal relationships between antecedent attributes and consequents. The aim of this paper is to develop a BRB system that complements the RiskMetrics WealthBench system for portfolio optimisation with nonlinear cash-flows and constraints. Two optimisation methods are presented to locate efficient portfolios under different constraints specified by the investors. Numerical studies demonstrate the effectiveness and efficiency of the proposed methodology.     

Random effects logistic regression model for ranking efficiency in data envelopment analysis

Ranking efficiency based on data envelopment analysis (DEA) results can be used for grouping decision-making units (DMUs). The resulting group membership can be partly related to the environmental characteristics of DMU, which are not used either as input or output. Utilizing the expert knowledge on super efficiency DEA results, we propose a multinomial Dirichlet regression model, which can be used for the purpose of selection of new projects. A case study is presented in the context of ranking analysis of new information technology commercialization projects. It is expected that our proposed approach can complement the DEA ranking results with environmental factors and at the same time it facilitates the prediction of efficiency of new DMUs with only given environmental characteristics.     

Parameter estimation in a condition-based maintenance model

A parameter estimation problem for a condition-based maintenance model is considered. We model a failing system that can be in a healthy or unhealthy operational state, or in a failure state. System deterioration is assumed to follow a hidden, three-state continuous time Markov process. Vector autoregressive data are obtained through condition monitoring at discrete time points, which gives partial information about the unobservable system state. Two kinds of data histories are considered: histories that end with observable system failure and histories that end when the system is suspended from operation but has not failed. Maximum likelihood estimates of the model parameters are obtained using the EM algorithm and a closed form expression for the pseudo-likelihood function is derived. Numerical results are provided which illustrate the estimation procedure.     

PLS分析与RBF神经网络耦合环境模型

鉴于城市大气环境质量受到诸多复杂因素影响,且各因素间存在多重相关性,本文将偏最小二乘(PLS)分析与人工神经网络径向基网络(RBF)耦合,建立偏最小二乘径向基神经网络模型(PLSRBF),应用于贵阳大气环境质量的检验和预测。实例表明:PLSRBF模型可对原多自变量模型进行降维简化,并可有效提取解释变量信息,防止信息丢失,且具有较强的拟合能力。     

A problem-structuring model for analyzing transportation–environment relationships

This study discusses a decision support framework that guides policy makers in their strategic transportation related decisions by using multi-methodology. For this purpose, a methodology for analyzing the effects of transportation policies on environment, society, economy, and energy is proposed. In the proposed methodology, a three-stage problem structuring model is developed. Initially, experts’ opinions are structured by using a cognitive map to determine the relationships between transportation and environmental concepts. Then a structural equation model (SEM) is constructed, based on the cognitive map, to quantify the relations among external transportation and environmental factors. Finally the results of the SEM model are used to evaluate the consequences of possible policies via scenario analysis. In this paper a pilot study that covers only one module of the whole framework, namely transportation–environment interaction module, is conducted to present the applicability and usefulness of the methodology. This pilot study also reveals the impacts of transportation policies on the environment. To achieve a sustainable transportation system, the extent of the relationships between transportation and the environment must be considered. The World Development Indicators developed by the World Bank are used for this purpose.     

Asymptotic risks of Viterbi segmentation

We consider the maximum likelihood (Viterbi) alignment of a hidden Markov model (HMM). In an HMM, the underlying Markov chain is usually hidden and the Viterbi alignment is often used as the estimate of it. This approach will be referred to as the Viterbi segmentation. The goodness of the Viterbi segmentation can be measured by several risks. In this paper, we prove the existence of asymptotic risks. Being independent of data, the asymptotic risks can be considered as the characteristics of the model that illustrate the long-run behavior of the Viterbi segmentation.     

隐马尔可夫模型在语音识别中的应用

隐马尔可夫模型 (HMM)的基本技术是语音识别中较为成功的算法 .主要是它具有较强的对时间序列结构的建模能力 .本文首先深入浅出地介绍了 HMM的基本技术和一个基于 HMM的孤立词语音识别系统的构成方法 ,其次 ,基于 HMM尚存有一些缺陷 ,造成语音识别能力较弱 ,为此本文又进一步阐述了语音识别应用中的几种改进的 HMM系统及目前的热点方法—— HMM与 ANN构成的混合网络     

A two-stage prognosis model in condition based maintenance

We often observe in practice that the life of a piece of production equipment can be divided into two stages. The first stage is referred to as the normal working stage where no significant deviation from the normal operating state is observed. The second stage is called the failure delay period, since a defect may be initiated, and progressively develop into an actual failure, i.e., the equipment is in a defective stage but still working during this stage. With the help of condition monitoring, hidden defects already present in the equipment may be detected, but for maintenance planning purposes, the prediction of the initiation point of the second stage, and more importantly, the residual life thereafter is important. This paper reports on the development of a probability model to predict the initiation point of the second stage and the remaining life based on available condition monitoring information. The method for model parameters estimation is discussed and applied to real data.     

An equilibrium model for urban transit assignment based on game theory

The urban public transport system is portrayed as a special commodity market where passenger is consumer, transit operator is producer and the special goods is the service for passenger’s trip. The generalized Nash equilibrium game is applied to describe how passengers adjust their route choices and trip modes. We present a market equilibrium model for urban public transport system as a series of mathematical programmings and equations, which is to describe both the competitions among different transit operators and the interactive influences among passengers. The proposed model can simultaneously predict how passengers choose their optimal routes and trip modes. An algorithm is designed to obtain the equilibrium solution. Finally, a simple numerical example is given and some conclusions are drawn.     

An S-Plus Implementation of Hidden Markov Models in Continuous Time

Abstract

Hidden Markov models (HMM) can be applied to the study of time varying unobserved categorical variables for which only indirect measurements are available. An S-Plus module to fit HMMs in continuous time to this type of longitudinal data is presented. Covariates affecting the transition intensities of the hidden Markov process or the conditional distribution of the measured response (given the hidden states of the process) are handled under a generalized regression framework. Users can provide C subroutines specifying the parameterization of the model to adapt the software to a wide variety of data types. HMM analysis using the S-Plus module is illustrated on a dataset from a prospective study of human papillomavirus infection in young women and on simulated data.     

On approximation of smoothing probabilities for hidden Markov models

We consider the smoothing probabilities of hidden Markov model (HMM). We show that under fairly general conditions for HMM, the exponential forgetting still holds, and the smoothing probabilities can be well approximated with the ones of double-sided HMM. This makes it possible to use ergodic theorems. As an application we consider the pointwise maximum a posteriori segmentation, and show that the corresponding risks converge.     

Filtering of a Multi-Dimension Stochastic Volatility Model

In this article, we study a stochastic volatility model for a class of risky assets. We assume that the volatilities of the assets are driven by a common state of economy, which is unobservable and represented by a hidden Markov chain. Under this hidden Markov model (HMM), we develop recursively computable filtering equations for certain functionals of the chain. Expectation maximization (EM) parameter estimation is then used. Applications to an optimal asset allocation problem with mean-variance utility are given.