Application of domain decomposition method and improved evolutionary algorithm in optimization of spiral inductors on silicon

The rising demand for low-cost radio frequency integrated circuits has generated an increasing interest in planar spiral inductors on silicon. However, several kinds of parasitic effects degrade the quality factor of inductors, so it is essential to choose optimal parameters to gain high Q values under some constraints. Thus nonlinear constrained optimization problem is formed. Calculating inductance with domain decomposition method can avoid tedious calculation for mutual inductance, and handling constrained optimization with stochastic ranking-based evolutionary algorithm can avoid determining optimal penalty coefficient. The paper tries to combine both methods and gains better results in less time. Experiments show that combining both methods can solve the nonlinear constrained optimization problem efficiently and successfully.