Research Of The Evolutionary Algorithm Of Digital Circuit Based On CGP

Evolvable hardware is evolvable hardware that is a simulation of the process ofbiological evolution, it is the combination of Evolutionary Algorithm and ProgrammableLogic Devices, according to the changes of the surrounding environment to adjust its systemfunction.With the adaptive, self-organizing, self-healing characteristics, it is widely used inthe electronic circuit system design. In recent years, with the development of the aerospaceindustry, the accuracy requirements for the design of electronic circuits is also increasing,which makes hardware design is also more and more complex, what comes next is theconvergence of electronic circuit design appears slow, difficult to converge, the longevolution time, difficult problems such as evolution of a large circuit scale and so on. How toeffectively solve these problems is an urgent question for many researchers need to be solved.Evolutionary design of circuit is an important application in the field of EvolvableHardware, it bases on the Programmable Logic devices and searches the novel circuitstructure by the Evolutionary Algorithm for the automatic design of complex circuit. As thetheoretical basis of Evolvable Hardware, Evolutionary Algorithm plays a decisive role in thedevelopment of the study on the performance of Evolvable Hardware. In this paper, We beginwith the design of digital circuits of evolutionary circuit design, with Cartesian GeneticProgramming (CGP) as the basis algorithm, according to the existing in the process ofevolution lost potential solutions and correct fitness stalling effect of CGP, it is improved inthis paper, put forward a kind of does not limit the external output of adaptive Cartesiangenetic programming, integrating the thinking of hierarchical evolutionary, the improvedalgorithm is used for the evolution of complex digital circuit design and make the adder as anexample for the successful completion of the evolutionary design of circuit. The researchcontents of this paper are as follows:（1）First we introduce the working principles and realization process of EvolvableHardware, and the four main Evolutionary Algorithm used by Evolvable Hardware wereanalysis,then make a comparison of their similarities and differences, and finallyintroduction the evolution of the physical implementation of hardware support- Programmable Logic Devices.（2） This paper introduces the principle of the Cartesian Genetic Programming, itscoding method, fitness evaluation and evolution strategy in detail.（3） During the evolutionary process of Cartesian Genetic Programming it is easy tolose the right potential solutions of the questions, a new encoding method is put forwardwhich external output is not limited; This paper Research and analysis the fitness stallingeffect of Cartesian Genetic Programming during the evolutionary process, imitationmathematical statistics "variance" concept defines the concept of "population fitness variance"and proposed a kind of "population fitness variance" for the variation standard adaptiveCartesian genetic Programming. Based on the improved Cartesian GeneticProgramming,integrating the progressive thinking of―simple set of complex, smallstructures big‖, using hierarchical evolutionary strategy, the circuit design methods was usedfor complex circuit design, and the addition as an example of the successful completion of thecircuit design of the evolution.