Global optimization and design for microelectromechanical systems devices based on simulated annealing

Abstract

In this paper, we propose a new method of global optimal design with simulated annealing (SA) for microelectromechanical systems (MEMS) devices. The optimal design of MEMS devices, with a microgyroscope as our device example, has been carried out to find the maximum sensitivity satisfying constraints imposed by functional and geometrical constraints. The optimization algorithm (SA) used is essentially an iterative random search procedure with adaptive moves along the coordinate directions. It permits downhill moves under the control of a probabilistic criterion, thus tending to avoid the first local maxima encountered. The optimization results are verified and validated with the finite element method (FEM) and the boundary element method (BEM) IntelliSuite™ results and measured data. When compared with Microsoft Excel Solver™ (generalized reduced gradient algorithm), our SA-based optimization approach exhibits promising superiority, and finds the global solution easily. There is also good agreement among the numerical computation results generated by the SA algorithm, the simulation results generated by IntelliSuite™, and the measured data.