Fast Multi-Dimensional Feed Optimization on Reflector Antennas Using a Reference Design

We present a fast method for optimizing a feed antenna on a reflector system, leveraging a known reference reflector system design. The method uses the optimum feed parameters of the reference reflector, determined by a specific goal function, as the starting point of the optimization. This ensures exploration in a favorable region of the parameter space. Difference patterns between the two reflectors are utilized, in a Characteristic Basis Function Pattern (CBFP) interpolation framework, to estimate the radiation patterns of the desired reflector system within the design region, requiring a minimum amount of electromagnetic simulations. A practical example within a high-dimensional parameter space is provided to demonstrate the power of the method.