Analog Circuits Design Using Ant Colony Optimization

In this paper, we present, Ant Colony Optimization (ACO) algorithm, as a tool to design analog circuits, for given output specifications. ACO is a swarm intelligence algorithm, which was first proposed in early nineties, to solve the problems of combinatorial optimization. In this paper, a modified version of this algorithm, (for continuous domain) has been implemented to optimize transistor sizes. In this work, widths of transistors are found for three analog circuits, to achieve the given specifications. Also, ACO has been tested on an Analog to Digital (ADC) Converter. Performance of the ADC is improved by optimizing the widths of the transistors, to minimize the error in the output of the ADC. In the past, one of the most popular evolutionary algorithms, Genetic Algorithm (GA) has been found to be effective in optimizing transistor sizes. Therefore, to examine the solutions achieved by ACO, all the circuits are optimized by GA also. Results show that ACO is better than GA in finding transistor sizes. Also, ACO takes less time in optimization process. In this work, Perl code of algorithms has been coupled with HSPICE to do circuit simulations. All circuits are simulated using BSIM3v3 MOSFET models in 0.13μm, 0.18μm or 0.35μm CMOS processes.