Searching Multisolutions in Synthesis of High-Order Filters and Multiplexers Aided With Multiobjective Optimization

This article proposes a synthesis framework for high-order filters and multiplexers based on an algorithm named multiobjective-aided cooperative coevolution differential evolution (MCCDE), which can synthesize multiple solutions for customized topologies. The topology synthesis is first transformed into a multiobjective optimization problem (MOP), to realize the global search for multiple potential solutions and to avoid rapidly falling into a local minima. In the optimization process, to address complicated high-order filtering networks, a cooperative co-evolution algorithm embedded with the parallel computing technique is utilized to escape from local minima and accelerate the convergence. The Levenberg-Marquardt (LM) algorithm is then used to implement the local search for more accurate solutions when taking the global potential solutions as the initial values. Synthesis examples of a 14th-order filter and a 20th-order multiplexer are provided to demonstrate the effectiveness and efficiency of the proposed synthesis theory for multisolution high-order networks. In the end, a 10th-order dielectric waveguide filter is fabricated and measured to verify the proposed approach.