In this thesis, we furthur investigate stochastic comparisons of the largest or-der statistics from multiple-outlier exponential models with respect to likelihood ratio and hazard rate orders. Under some specified assumption, two sufficient conditions are provided for the likelihood ratio and hazard rate orders to hold between the largest order statistics arising from two multiple-outlier exponential models, respectively, we also extend the results to the proportional hazard rate models. In addition, the results established here strengthen and generalize some of the results known in the literature. Finally, some numerical examples are given to demonstrate main results obtained here. |