Challenging the evolutionary strategy for synthesis of analogue computational circuits

Abstract

There are very few reports in the past on applications of Evolutionary Strategy (ES) towards the synthesis of analogue circuits. Moreover, even fewer reports are on the synthesis of computational circuits. Last fact is mainly due to the dif-ficulty in designing of the complex nonlinear functions that these circuits perform. In this paper, the evolving power of the ES is challenged to design four computational circuits: cube root, cubing, square root and squaring functions. The synthesis succeeded due to the usage of oscillating length genotype strategy and the substructure reuse. The approach is characterized by its simplicity and represents one of the first attempts of application of ES towards the synthesis of “QR” circuits. The obtained experimental results significantly exceed the results published before in terms of the circuit quality, economy in components and computing resources utilized, revealing the great potential of the technique pro-posed to design large scale analog circuits.