On parametric uncertainty modeling for real parameter variations

A procedure for obtaining an uncertainty model for real uncertain parameter problems in which the uncertain parameters can be represented in a multilinear form is presented. The procedure is formulated such that the resulting uncertainty model is minimal relative to a given state-space realization of the system. The approach is demonstrated for a multivariable third-order example problem having four uncertain parameters. Although two of the uncertain parameters entered into the given model as squared terms and as fractions, they were easily modeled within the multilinear framework