Analog circuit topology synthesis by means of evolutionary computation. (April 2019)

• Record Type: Journal Article
• Authors: Rojec, Žiga; Bűrmen, Árpád; Fajfar, Iztok
• Is Part Of: Engineering applications of artificial intelligence. Volume 80(2019)
• Page Start: 48
• Page End: 65
• Publication Date: 2019-04
• Subjects: Evolutionary algorithms -- Multi-objective optimization -- Hybrid optimization -- Automated circuit synthesis -- Analog circuits; Engineering -- Data processing -- Periodicals; Artificial intelligence -- Periodicals; Expert systems (Computer science) -- Periodicals; Ingénierie -- Informatique -- Périodiques; Intelligence artificielle -- Périodiques; Systèmes experts (Informatique) -- Périodiques; Artificial intelligence; Engineering -- Data processing; Expert systems (Computer science); Periodicals; 620.00285
• DOI: 10.1016/j.engappai.2019.01.012
• Languages: English
• ISSNs: 0952-1976

Abstract

Abstract: Designing an analog electrical circuit requires substantial effort and time by a highly-skilled designer. Many tools have been devised to aid engineers in the design procedure, from simulators to parameter optimization tools, which largely assist in fine-tuning the parameter values of circuits with predetermined topologies. In this paper, we propose a new approach to circuit optimization that is capable of searching not only for the parameter but also the topological space. Starting from a high-level statement describing a desired circuit functionality, the proposed method can either improve a given circuit, or create a novel topology with given resources, based on the Darwinian theory of evolutionary selection. Our method uses a global parameter optimization method PSADE to speed up evolutionary convergence. We use the proposed algorithm (both single- and multi-objective) to successfully evolve passive, active, and BiCMOS circuits.