FPGA Design And Implementation Of Bacterial Foraging Optimization Algorithm

Bacterial Foraging Optimization(BFOA)is a biologically inspired swarm intelligence algorithm with excellent convergence performance and global optimization ability.The BFOA algorithm is usually implemented in software to optimize various multi-dimensional complex problems.With the development of science and technology,more researchers have begun to implement new intelligent algorithms on hardware.which can effectively improve the processing speed of the algorithms and make the integration of electronic products higher and higher.Therefore,in this paper,the BFOA algorithm is deployed and designed based on the hardware platform FPGA.and it is applied to the optimal processing of image segmentation threshold,which effectively improves the threshold optimization speed.This paper first introduces the principle and design process of the chemotaxis step,reproduction step and migration step of the BFOA algorithm.In view of the problem that the standard BFOA algorithm has a fixed chemotaxis step and the step size is too low.This results in poor performance of the algorithm.Combined with the random step size and the distance formula,the step size mechanism improvements have been made.It is verified on the Python platform that the improved BFOA algorithm has better convergence performance and ability to jump out of local extreme values than the standard BFOA algorithm.Secondly,the improved BFOA algorithm is divided into four modules,As a whole including chemotaxis module,replication module,migration module and storage module.Then design and realize the functions of each module step by step,and the overall algorithm is realized by state machine control loop.In the chemotaxis module,the flipping module,the walking module and the fitness value calculation module are implemented in parallel.As the trending module,the copying module and the migration module in the overall algorithm are pipelined,which improves the overall performance of the design scheme.The overall design uses the Verilog language to implement the BFOA algorithm on the platform of Vivado,and takes the Schaffer function as the optimization goal.Finally,the BFOA algorithm is applied to image segmentation,and the maximum entropy threshold segmentation method is used as the gray threshold evaluation function.According to the maximum entropy threshold method,the fitness value module is divided into a parameter module and a maximum entropy value calculation module.The random grayscale threshold is sent to the BFOA algorithm module for iterative update,and the entropy value is used as the evaluation standard for the random grayscale threshold.The optimization of the segmentation threshold is completed.And the performance test of the maximum entropy threshold segmentation method based on BFOA is carried out.According to the comparison of test time and results,it is proved that the algorithm can quickly and accurately locate the optimal segmentation threshold,and the accuracy of the application is verified at the same time.The BFOA algorithm based on FPGA in this paper has the advantages of portability and versatility.Since the design of the fitness value module can be flexibly transformed,the algorithm can be used in more complex problems.