Accurate permittivity measurement by means of inverse calculation based on genetic algorithms

The design of devices or systems in several microwave applications, such as communications, radar, or industrial heating, requires an accurate knowledge of the dielectric properties of the different materials employed. Direct methods of measurement of the complex permittivity can be quick, but sometimes they do not provide the required accuracy. Moreover, these methods need the sample to have a canonical shape (prisms, cylinders, etc.) and, when a nonhomogeneous material must be measured in order to obtain the permittivities of its elements, they obtain an effective permittivity, but are not able to measure the individual permittivities of the elements. Some work has been done on estimating the elemental permittivities in nonhomogeneous samples. We present an experimental/computational hybrid method for estimating the complex permittivity of homogeneous or nonhomogeneous materials with arbitrary shape by means of an inverse procedure.