FPGA implementation of a cellular univariate estimation of distribution algorithm and block-based neural network as an evolvable hardware

Abstract

Abstract-This paper pr esents a har dwar e implementation of evolvable block-based neur al networ k (BBNN) amd a kind of EDAs called cellular compact genetic algor ithm (CCGA) in FPGA. The CCGA and BBNN have cellular -like and ar r ay-like str uctur es which ar e suitable for har dwar e implementation. The implemented har dwar e demonstr ates the completely intr insic online evolution in har dwar e without softwar e r unning on micr opr ocessor s. This wor k contr ibutes to the field of evolvable har dwar e by pr oposing CCGA and a layer -based ar chitectur e to an integr ation of BBNN and CCGA as a kind of evolvable har dwar e.In addition, the pr oposed CCGA efficiently solves the scalable issues by scaling up to the size of BBNN. The pr esented appr oach demonstr ates a new kind of evolvable har dwar e.