A gradient-based optimization technique employing determinants for the synthesis of microwave coupled filters

A procedure is presented whereby cross-coupled filters can be synthesized using a gradient-based optimization technique employing determinants. The cost of function formulated depends on a minimum set of characteristics that can completely describe the desired response. An analytical gradient is also derived for the error function enhancing the efficiency of the optimization process. The proposed algorithm is less vulnerable to local minima and proves more robust compared to other optimization techniques. Examples demonstrating the efficiency of the proposed algorithm are provided.