A simulation-based hybrid optimization technique for analog and digital integrated circuits design automation

In this paper, a simulation-based optimization technique for integrated circuit (IC) design automation is presented. Based on biological inspired techniques, numerical deterministic methods and circuit simulator, the hybrid optimization methodology is developed for IC design. Two different circuits, low noise amplifier (LNA) and static random access memory (SRAM) are analyzed with our technique. Considering LNA IC, we simultaneously evaluate specifications including S parameters, K factor, noise figure, and input third-order intercept point in the optimization process. If the simulated results meet the aforementioned constraints, we output the optimized parameters. Otherwise, one of evolutionary algorithms will enable us to search solution globally; simultaneously, numerical optimization methods solve the solutions according to the results of global search. A circuit simulator calculates the circuit's characteristics with the updated configurations, and the developed prototype will evaluate newer results until all specifications are matched. Similarly, by considering the static noise margin (SNM), optimal device's channel length and supply voltage for SRAM cells with different circuit types could be tuned. Our preliminary results confirm the robustness and efficiency of the proposed simulationbased optimization technique.