Design of Electrooptic Modulators Using a Multiobjective Optimization Approach

This paper presents preliminary results of a multiobjective approach for optimization of design parameters of Mach-Zehnder-based lithium niobate modulators. This process uses a genetic algorithm to iteratively refine candidate sets of parameters, and the characterization of the modulators is performed by the finite-element method. Test cases include optimization of a conventional Mach-Zehnder modulator and a variation of this kind of device using additional floating electrodes. The set of characteristics includes parameters such as the characteristic impedance, the effective index of the microwave, and the half-wave voltage.