ABSTRACT
Different synthesis approaches such as knowledge-based approach, equation-based approach, and simulation-based approach are widely used for synthesis of analog circuits. Each of these approaches has its own advantages and disadvantages. The most popularly used simulation-based approach using simulation tools provides the most accurate solution, but it consumes huge time which effect the time-to-market of the system-on-chip (SoC). Performance-based macromodeling plays a vital role in analog circuit synthesis and optimization. Macromodels of analog circuits are essential for the automatic and fast synthesis of analog circuits. These synthesized circuits are the building blocks of SoCs. Macromodels need to be developed very fast, even for the complex circuits consisting of a larger number of transistors without affecting the accuracy of the original circuit.
Obtaining the input parameters of the circuits to operate in the given design space or obtaining the optimized output parameters of the circuits is the major bottleneck for the circuit designers. With the increase in the number of design variables, the synthesis of analog circuit in large design space is getting more and more difficult. So various macromodels such as symbolic models, posynomial models, and neural models have been reported by researchers in the literature. These macromodels are used in place of SPICE simulation in the circuit synthesis flow. This chapter is mainly focused on review of various types of macromodeling techniques used for analog integrated circuits.
