APPLICATION OF OPTIMIZATION METHODS AND ALGORITHMS FOR MEMS SWITCH MODELING

The paper discusses the modernized development of the design optimization method for MEMS switches with metal-to-metal resistive contacts to improve its performance characteristics, such as the reflection coefficient and the on-state transfer coefficient, and the off-state isolation. To achieve this goal, the physical parameters of the design are considered, namely the length and width of the lower metal contact pad of the transmission line, which are changed simultaneously. Three methods are used in the optimization process: the quasi-Newton method, the genetic algorithm, and the Hooke-Jeeves method. These methods allow us to explore different approaches to finding the optimal design parameters and compare their effectiveness.