Random group associated with differential evolution improves scalability of analog circuit design

Abstract

The scalability problem is an important task to evolutionary analog circuit design. In this paper, we aimed at raising the search efficiency of evolutionary algorithm to improve the scalability of evolutionary analog circuit design. We made an analysis and found out that all the components in circuits are related. After encoding, the genes belong to the chromosome are correlation and non-separable. Evolutionary analog circuit design is a non-separable problem. A random group method associated with differential evolution (DE) is proposed in this paper. In the experiment section, standard DE and DE with random group method (RGDE) are used for the scale experiments. The results show that random group method associated with DE can raise the search efficiency. RGDE is conducive to improving the scalability of analog circuit evolutionary design.