Research And Application Of Hardware Implementation Of The Genetic Algorithm Based On FPGA

Nowadays, with the rapid development of science and technology, the geneticalgorithm has been applied in more and more areas and been put forward higherrequirements. The study of the implementation and performance of the genetic algorithmin order to enable it meets the higher requirements. The current research on geneticalgorithm is more concentrated in the aspect of software implementation than hardware.The software implementation of the genetic algorithm would slow down the speed ofcalculation because of the nature of the genetic algorithm is parallel, however, hardwareimplementation is ideal for it.This article introduces the basic principles of the genetic algorithm and steps ofimplementation of each Genetic operators. Various implementation ways of each geneticoperators are listed and analyzed. The situation of hardware implementation of thegenetic algorithm is introduced. Advantages and disadvantages of various hardwareimplementations are point out also. So we know where needed to be further improved inthose implementations. And then the article introduces the hardware implementationplatform of the algorithm that includes devices, programming languages and simulationsoftware and so on.The man work and innovation points of the article:(1) A new pipeline is designed to improve the speed of the algorithm.(2) The article designs the overall structure of the algorithm; Two parallel circuitsare used in the algorithm; Steps of the algorithm are realized by the module, and themare designed to be suitable for hardware implementation; Deterministic CrowdingGenetic Algorithm is added to enable the algorithm deal with multi-peak functions; Inorder to overcome the prematurity problem in the niche method, the mutation probabilityis adaptively adjusted, so the searching performance of the algorithm is further improved.The experimental results indicate that the hardware implementation of this geneticalgorithm could find all the local optimal solutions and the global optimal solution; thespeed of the algorithm is two orders of magnitude or above faster than the softwareimplementation.(3) The hardware implementation of the improved genetic algorithm is used totuning parameters of the PID controller to make an intelligent PID controller, and thecontroller is applied in a resistance furnace temperature control system. Theexperimental results confirm its excellent performance.At the end of the article, the summary has been carried on. Some deficiencies thatneed further improvement are puts forward.