Circuit Automation Design Based On Genetic Algorithm

The automatic design of circuits is an important branch of evolutionary hardware.This circuit automatically adjusts the circuit structure to adapt to environmental changes without manual intervention,and has the characteristics of self-organization and self-repair.In recent years,due to the advantages of evolutionary hardware circuits that consume less logic resources and lower power consumption,the use of evolutionary algorithms for automatic circuit design has become a research hotspot.In addition,programmable logic devices have the advantage of repeatable configuration of internal structure functions,making it highly feasible to implement evolutionary algorithms on hardware platforms.As an important branch of evolutionary algorithm,genetic algorithm has the characteristics of good selection,mutation and search for optimal solution.Combining it with FPGA hardware language for parallel processing can greatly improve the convergence speed of the algorithm.Therefore,this paper applies the modular design ideas of the state machine to the genetic algorithm design,and transplants it to the automatic design of the combinational logic circuit,which has achieved good results.The main work of this paper is as follows:Firstly,the basic principle and implementation technology of genetic algorithm are analyzed systematically,and each module of genetic algorithm is introduced in detail.Then the hardware implementation platform of this design,FPGA(field programmable gate array),its simulation software and programming language,is introduced.Then the natural parallelism of genetic algorithm is applied to the FPGA platform.According to the top-down design idea of FPGA,the algorithm is divided into modules.The division idea is based on the modular design of finite state machine.In this paper,a normalized system design method is proposed.Through the medium-sized integrated counter,the state centralized combination logic part is designed,and the time sequence logic part is designed through the counter The validity of the method is verified by experiments.Then the function of each module of the algorithm is introduced in detail.In coding mode,gene expression programming algorithm is used,and logic gate and input signal are combined into logic block,which greatly shortens the length of binary coding.Expression tree is used to translate expression tree into logic circuit from bottom to top.At the same time,the crossover operator of two-point crossover is used to exchange some genes,and then the crossover operator and mutation operator are used to update some genes to ensure the diversity of the population and improve the local search ability of genetic algorithm.Finally,Verilog HDL language is used to write codes to realize the functions of each module.In the vivado 2018.3 development integration environment,the syntax analysis,inspection,function simulation,comprehensive layout and wiring of each module are carried out respectively,and the verification of the whole design is completed.The experimental results show that the modular design idea of this paper has been verified in the genetic algorithm based on FPGA,and the structure is simple,the convergence speed has been improved,and it has a certain practicality,which makes the basic preparations for the application of the algorithm in more complex and higher-speed scenes.