Investigation of particle swarm optimization for dynamic reconfiguration of field-programmable analog circuits

The sensor electronics and mixed signal processing circuits are sensitive to aging, temperature distribution inside the chip, manufacturing tolerance, or additional factors of the external environment. These influences can't be effectively considered in the design phase. Some applications requiring costly and time consuming trimming, but still can't deal with all the mentioned problems. The goal of this paper is to investigate the particle swarm optimization (PSO) as an alternative to genetic algorithm (GA) for static and dynamic reconfiguration (e.g. compensating the mentioned influences) of field programmable analog scalable device array (FPADA) (Tawdross and Konig, 2005; Lakshmanan and Konig, 2005). In order to have a reliable circuit with a predictable performing, the main topology of the hardware (HW) is chosen by human or by other selection topology. We adopt the PSO to design an electronic system (operational amplifier) which is a new application to the PSO. We compare the results of the PSO with our previous work in Tawdross et al. (2005) and Tawdross and Konig (2005), which used the GA to investigate the ability of the PSO for HW design and reconfiguring the FPADA. The results show that the PSO succeeds to find a better, faster and easier to implement solution than the GA.