Understanding the shape of the mixture failure rate (with engineering and demographic applications)

Mixtures of distributions are usually effectively used for modelling heterogeneity. It is well known that mixtures of DFR distributions are always DFR. On the other hand, mixtures of IFR distributions can decrease, at least in some intervals of time. As IFR distributions often model lifetimes governed by ageing processes, the operation of mixing can dramatically change the pattern of ageing. Therefore, the study of the shape of the observed (mixture) failure rate in a heterogeneous setting is important in many applications. We study discrete and continuous mixtures, obtain conditions for the mixture failure rate to tend to the failure rate of the strongest populations and describe asymptotic behaviour as t→∞. Some demographic and engineering examples are considered. The corresponding inverse problem is discussed. Copyright © 2009 John Wiley & Sons, Ltd.     

Mixture of Weibull distributions—parametric characterization of failure rate function

A mixture distribution involving two two-parameter Weibull distributions is characterized by five parameters. The shape of the failure rate function for the mixture depends on the parameter values and can be one of eight distinct shapes. In this paper we present a complete parametric characterization of the shapes for the failure rate function for the mixture. © 1998 John Wiley & Sons, Ltd.     

A two-parameter of weighted exponential distributions

In this paper, a class of distributions called the two-parameter weighted exponential distributions is introduced (TWE). This new class of distributions generalizes the ones-weighted exponential distributions (WE)-proposed by Gupta and Kundu (2009). The main properties of this new class of distributions are investigated. Several statistical properties and statistical inferences are then obtained and studied. Two real data sets of which one is a right censored data set were analyzed, and it is shown that in both two cases our model fits much better than WE or some other existing models.     

Conditional laplace transforms for bayesian nonparametric inference in reliabity theory

In order to apply nonparametric meyhods to reliability problems, it is desibrable to have available priors over a broad class of survival distributions.In the paper, this is achieved by taking the failure rate function to be the sum oof a nonnegative stochastic process with increasing sample patjhs and a process with decreasing sample paths. This approach produces a prior which chooses an absolutely survival distribution that can have an IFR, DFR, or U-shapped failure rate. Posterior Laplace transforms of the failure rate are obtained based on survival data allows censoring. Bayes estimates of the failure rate as well as the lifetime distribution are then calculated from these posterior Laplace transforms. This approach is also applied to a competing risks model and the proportional hazards model of Cox.     

On a new multivariate IFR ageing notion based on the standard construction

Many criteria of ageing for random variables or vectors have been proposed in the literature over many years. For instance, a random variable is increasing in failure rate (IFR) if, and only if, it can be ordered with an exponentially distributed random variable in the classical univariate convex transform order. A new multivariate generalization of the convex transform order has recently been proposed in the literature. In this work, we propose a new multivariate IFR notion for multivariate distributions based on comparisons in this new order with a properly defined exponentially distributed random vector. Properties, applications, and illustrations of this new notion are given as well. Copyright © 2016 John Wiley & Sons, Ltd.     

A lifetime model with increasing failure rate

This paper deals with a new two-parameter lifetime distribution with increasing failure rate. This distribution is constructed as a distribution of a random sum of independent exponential random variables when the sample size has a zero truncated binomial distribution. Various statistical properties of the distribution are derived. We estimate the parameters by maximum likelihood and obtain the Fisher information matrix. Simulation studies show the performance of the estimators. Also, estimation of the parameters is considered in the presence of censoring. A real data set is analyzed for illustrative purposes and it is noted that the distribution is a good competitor to the gamma, Weibull, exponentiated exponential, weighted exponential and Poisson-exponential distributions for this data set.     

Comparison of maintenance policies with monotone failure rate distributions

We compare different maintenance policies assuming that the system lifetime has either increasing failure rate (IFR), or decreasing failure rate (DFR) distribution. We show that these assumptions yield strongly stochastic comparisons between the process' intensities. This yields weaker stochastic comparison between the processes than the stochastic comparison that hold when the system lifetime is new better than used (NBU) or new worse than used (NWU). Copyright © 2002 John Wiley & Sons, Ltd.     

A hybrid Pareto model for asymmetric fat-tailed data: the univariate case

Density estimators that can adapt to asymmetric heavy tails are required in many applications such as finance and insurance. Extreme value theory (EVT) has developed principled methods based on asymptotic results to estimate the tails of most distributions. However, the finite sample approximation might introduce a severe bias in many cases. Moreover, the full range of the distribution is often needed, not only the tail area. On the other hand, non-parametric methods, while being powerful where data are abundant, fail to extrapolate properly in the tail area. We put forward a non-parametric density estimator that brings together the strengths of non-parametric density estimation and of EVT. A hybrid Pareto distribution that can be used in a mixture model is proposed to extend the generalized Pareto (GP) to the whole real axis. Experiments on simulated data show the following. On one hand, the mixture of hybrid Paretos converges faster in terms of log-likelihood and provides good estimates of the tail of the distributions when compared with other density estimators including the GP distribution. On the other hand, the mixture of hybrid Paretos offers an alternate way to estimate the tail index which is comparable to the one estimated with the standard GP methodology. The mixture of hybrids is also evaluated on the Danish fire insurance data set.      

Statistical analysis of progressive stress accelerated life test for the product of two-parameter Laplace BS fatigue life distribution under inverse power law model

Based on the product of two-parameter Laplace Birnbaum-Saunders fatigue life distribution, its failure distribution mode is theoretically derived under the progressive stress accelerated life test with inverse power law model, and then three-parameter generalized Laplace Birnbaum-Saunders fatigue life distribution is introduced. The basic properties of three-parameter generalized Laplace Birnbaum-Saunders fatigue life distribution are analyzed, and the image characteristics of its density function, failure rate function and average failure rate function are investigated. Meanwhile, the point estimate method is given for three parameters, and then the point estimates of parameters are obtained for the product of two-parameter Laplace Birnbaum-Saunders fatigue life distribution under the progressive stress accelerated life test with inverse power law model. In addition, the practical example and simulation examples are illustrated to show the feasibility of the proposed method.     

A reliability application of a mixture of inverse gaussian distributions

A mixture of inverse Gaussian distributions (IGDs) is examined as a model for the lifetime of components. The components differ in one of three ways: in their initial quality, rate of wear, or variability of wear. These three cases are well represented by the parameters of the IGD model. The mechanistic interpretation of the IGD as the first passage time of Brownian motion with positive drift is adopted. The parameters considered are either dichotomous or continuous random variables. Parameter estimation is also examined for these two cases. The model seems to be most appropriate when the single IGD model fails due to heterogeneity of the initial component quality.     

A new approach for parameter estimation of finite Weibull mixture distributions for reliability modeling

The aim of this paper is to model lifetime data for systems that have failure modes by using the finite mixture of Weibull distributions. It involves estimating of the unknown parameters which is an important task in statistics, especially in life testing and reliability analysis. The proposed approach depends on different methods that will be used to develop the estimates such as MLE through the EM algorithm. In addition, Bayesian estimations will be investigated and some other extensions such as Graphic, Non-Linear Median Rank Regression and Monte Carlo simulation methods can be used to model the system under consideration. A numerical application will be used through the proposed approach. This paper also presents a comparison of the fitted probability density functions, reliability functions and hazard functions of the 3-parameter Weibull and Weibull mixture distributions using the proposed approach and other conventional methods which characterize the distribution of failure times for the system components. GOF is used to determine the best distribution for modeling lifetime data, the priority will be for the proposed approach which has more accurate parameter estimates.     

TFR模型序加试验下WEIBULL分布产品寿命的统计分析

本文针对损伤效率(TFR)模型,首次提出将步加试验推广至序加试验,给出了两参数Weibull分布参数的极大似然估计.     

一个新的具有单调降失效率的寿命分布

本文由Pareto分布和Logarithmic分布"混合"生成两参数具有单调降失效率的新型寿命分布,研究了该分布的矩、熵、失效率函数、平均剩余寿命和参数的极大似然估计,应用EM算法求参数的极大似然估计,进行了数值模拟.     

A Note on the Multivariate Normal Hazard

The ISO* property of noncentrality parameters is derived for the expected value of an ISO* function of independent nonnegative two-parameter compound Poisson random variables and is then applied to unbiasedness of tests and monotonicity of power functions of tests in an order-restricted hypothesis testing problem for the noncentrality parameter. The ISO* property and the Schur convexity are also studied for a class of two-parameter distributions which has the additive property (i.e., is closed under convolution) and contains the one-parameter family with the semigroup property as a special case.     

On an inverse problem in mixture failure rates modelling

Mixtures of decreasing failure rate (DFR) distributions are always DFR. It turns out that very often mixtures of increasing failure rate distributions can decrease or show even more complicated patterns of dependence on time. For studying this and other relevant effects two simple models of mixing with additive and multiplicative failure rates are considered. It is shown that for these models an inverse problem can be solved, which means that given an arbitrary shape of the mixture failure rate and a mixing distribution, the failure rate for a governing distribution can be uniquely obtained. Some examples are considered where this operation can be performed explicitly. Possible generalizations are discussed. Copyright © 2001 John Wiley & Sons, Ltd.     

On modeling failure and repair times in stochastic models of manufacturing systems using generalized exponential distributions

Failures of machines have a significant effect on the behavior of manufacturing systems. As a result it is important to model this phenomenon. Many queueing models of manufacturing systems do incorporate the unreliability of the machines. Most models assume that the times to failure and the times to repair of each machine are exponentially distributed (or geometrically distributed in the case of discrete-time models). However, exponential distributions do not always accurately represent actual distributions encountered in real manufacturing systems. In this paper, we propose to model failure and repair time distributions bygeneralized exponential (GE) distributions (orgeneralized geometric distributions in the case of a discretetime model). The GE distribution can be used to approximate distributions with any coefficient of variation greater than one. The main contribution of the paper is to show that queueing models in which failure and repair times are represented by GE distributions can be analyzed with the same complexity as if these distributions were exponential. Indeed, we show that failures and repair times represented by GE distributions can (under certain assumptions) be equivalently represented by exponential distributions.This work was performed while the author was visiting the Laboratory for Manufacturing and Productivity, Massachusetts Institute of Technology, Cambridge, MA 02139, USA.     

一般寿命分布和定时截尾的Bayes变量抽样方案

林（1994）研究了指数分布和定时截尾的变量抽样方案．本文将讨论一般寿命分布和定时截尾的一次抽样方案．在多项式损失函数的假设下,我们讨论了Weibull分布、双参数指数分布和-分布三种情形,并着重讨论Weibull分布的情形．本文还提出了一个可用于近似地确定最优抽样方案的有报算法,并且进行了灵敏度分析,还同林较早的模型（1990,1994）做了比较．     

两参数指数分布产品全样本场合下步进应力加速寿命试验损伤失效率模型下的统计分析

本文给出了两参数指数分布产品全样本场合下步进应力加速寿命试验损伤失效率模型下参数的极大似然估计和拟矩估计,并通过Monte-Carlo模拟说明本文方法的可行性.     

On exponential scores statistics for testing against positive aging

It has been shown in this paper that a test proposed by Barlow and Doksum (1972) based on the exponential scores statistic for testing exponentiality against increasing failure rate distributions is consistent for the much wider class of harmonic new better than used in expectation distributions.