The US Navy has increased interest in the reliability of aircraft launching and recovery equipment. Data is readily available and failure time distributions can be estimated; however, the aircraft equipment will not be operated with the same stress profile in the future as the data provided. In fact, the US Navy will increase the stress profile on the equipment by incorporating heavier aircraft into the fleet, while downsizing the lighter aircrafts. This creates an uncertain stress profile the aircraft carrier systems will be subjected to. Since the composition of the fleet is uncertain, determining reliability and component redundancy and/or replacement is difficult. Thus, new models and optimization algorithms are proposed involving data analysis at the component-level based on Weibull shape parameters modeled after using a general log-linear model based on the mean and variance of critical stress measures in a changing environment, and Weibull shape parameters modeled using a general log-linear model based on the distributional form of critical stress measures in a changing environment. Traditional system reliability considers a set of failure data which is analyzed to estimate a failure time distribution. This failure time distribution can be utilized to estimate reliability at some point in time. This thesis pertains to design problems with a probabilistic future stress profile, but using models based upon the current failure data. Since a future stress profile can be probabilistic and distinctly different, the traditional system reliability model will be unable to estimate future reliability from the existing failure data. Instead an estimate of the future failure time distribution must be made utilizing accelerated life concepts, and the optimal component reliability becomes difficult to determine. Depending on the level of usage, the optimal component redundancy might change. This research project tries to develop a heuristic for system reliability optimization considering a probabilistic future stress profile in which the stresses can increase to different levels. A failure time distribution will be determined for each system component as a function of usage stress distribution. The component models are then assembled into a system model. This system model will test different composition of fleet data based upon different probabilities. Although these probabilities are ambiguous it is certain that the stress profiles will increase. This system model was tested to see what optimal preventative maintenance or component replacement can be done in the present so that the unknown future stress profile will not cause high costs in maintenance and replacement parts.