Search

This paper develops a Bayesian analysis of the scale and shape parameters in the Weibull distribution and the corresponding reliability function with respect to the usual life-testing procedures. For the scale parameter θ, Bayesian estimates of θ and reliability are obtained for the uniform, exponential, and inverted gamma prior probability densities. BHATTACHARYA'S results [J. Am. Stat. Assn. 62, 48–62(1967)] for the one-parameter exponential life-testing distribution are reduced to a special case of these rewards. The paper develops a fully Bayesian analysis of both the scale .and shape parameters θ and ξ by independent prior distributions; since in the latter case, analytical tractability is not possible, Bayesian estimates are obtained through a conjunction of a Monte Carlo simulation and numerical-integration techniques. In both cases, the paper carries out a computer simulation and makes a comparison between the Bayesian and the corresponding minimum-variance unbiased, or maximum likelihood, estimates. As expected, the Bayesian estimates are superior. [ABSTRACT FROM AUTHOR]

This paper describes the use of multiple ranking procedures to analyze data generated from computer simulation experiments with models of management systems. After outlining the rationale for the use of multiple ranking procedures with computer simulation experiments and defining some basic terminology, we examine several specific multiple ranking procedures. Careful attention is given to the assumptions underlying the different multiple ranking procedures and the extent to which these assumptions are satisfied by the data generated by simulation experiments. An example model is included to illustrate the applicability of multiple ranking procedures to simulation experiments in management science. [ABSTRACT FROM AUTHOR]

The technique presented permits the experimenter to eliminate, if possible, a large number of the design points of an experiment. It also permits him to sequentially converge, if necessary, to an entire set of design points or to eliminate design points in which there is little interest. The technique is very flexible with regard to augmentation of levels and/or factors at various stages of the investigation. Thus, it provides a means of designing and augmenting experiments which is particularly suitable to simulation experiments. Furthermore, the analysis of the data at each step in the sequence is performed using the conventional analysis-of-variance techniques. [ABSTRACT FROM AUTHOR]