A new method for optimization of analog integrated circuits using Pareto-based multi-objective genetic algorithm

This paper presents a new method for design and optimization of analog integrated circuits based on Pareto-based Multi-Objective Genetic Algorithm (MOGA). The efficiency of the method is evaluated by using benchmark problems and compared with other MOGA algorithms. The method is implemented and used to optimize a telescopic cascode Op-Amp. Here, transistor sizes, compensation capacitor and bias voltages are determined by Genetic Algorithm (GA). Moreover, the circuits that are formed, using the components of the determined values, are simulated with Hspice. The output parameters, such as gain, bandwidth, phase margin and power are extracted from the generated output file, and the area of chip is calculated separately. The extracted output parameters are used as cost functions for creating next generation in GA. Finally, a set of Pareto-front which satisfies the conditions of the problem is introduced. This enables the circuit designer to select the best solution from the set. This algorithm is implemented in Matlab and is simulated by using Hspice and TSMC 0.18um CMOS technology is employed in simulation.