Modeling and Optimization of Integration Processes under Severe Uncertainties

The integration process of engineering systems frequently involves high uncertainties, such as units' availability time, integration increment duration, testing costs, etc. In this work the uncertainty effect is investigated by three methods: statistical analysis under the assumption that the mean and variance of the underlying distributions are known, a Monte Carlo based method when the uncertain values can be bounded from below and above by known quantities, and information gap theory when only nominal values and relative weights regarding the unknown variables are given. The optimization is performed by dynamic programming. The modeling methods and the associated optimization are applied to a small tactical missile, examining and comparing the various strategies.