In-partition in-stage evolutionary mechanism by applying input and output decomposition
Yao Ru · Harbin Gongcheng Daxue Xuebao/Journal of Harbin Engineering University · 2015
To solve the problem of scalability of evolvable hardware,an in-partition in-stage parallel online evolution mechanism based on input and output decomposition is presented for evolving combinational logic circuit. According to the input and output decomposition strategy,the original combinational logic circuit is divided into several sub-circuits with fewer inputs and outputs. Each sub-circuit is assigned to a separate evolution area. This is constructed to realize the parallel evolution of sub-circuits. When the evolutions of some sub-circuits are completed,the corresponding areas of evolution can be used for the parallel evolution of any other sub-circuits with uncompleted evolution. After the evolution on all sub-circuits are successfully completed they are integrated into a top circuit.In the self-evolving system built by Xilinx Virtex-5 FX,the adder circuit,multiplier circuit and partial MCNC benchmark circuits are taken as the examples for demonstration. The experimental results showed that compared to the classical evolutionary method,this method may greatly reduce the evolution time and evolve a combinational logic circuit with 21 inputs at most.