Microcosm algorithm hardware platform realizing method based on FPGA

Abstract

The invention discloses a field programmable gate array-based small-world algorithm hardware system. Based on the principle of the small-world algorithm, it consists of two levels of individual optimization and global optimization, and is connected by individual information modules in the middle. The individual optimization part is composed of a random number generator, a position reorganization module, a fitness calculation module and a comparison module connected in sequence. Complete the individual neighborhood search, fitness calculation and selection process. The global optimization part is composed of individual fitness comparison module and system scheduling module connected sequentially. The individual fitness comparison module selects the best individual among all individuals, and supplies the system scheduling module to judge whether the termination condition is met, and whether to initiate the next generation search, etc. This method has simple hardware implementation, high parallelism, can effectively overcome the local extremum problem, retain the diversity of the solution space, and overcome the deception problem more effectively to a certain extent, and has faster convergence speed and stability.

Description

technical field [0001] The invention belongs to the application field of artificial intelligence optimization algorithms, and in particular relates to a hardware platform implementation method of small-world optimization algorithms on FPGA. Background technique [0002] In scientific research and engineering practice, it is often necessary to solve various optimization problems, which requires the use of various optimization methods. At present, optimization methods mainly include classical optimization methods and heuristic optimization algorithms. Classical optimization methods mainly include linear programming, gradient method, Newton method, interior point method, etc. Heuristic algorithm is another important branch, including genetic algorithm, artificial neural network algorithm, ant colony algorithm, particle swarm algorithm and simulated annealing algorithm. Because the classical optimization algorithm has more requirements on the nature of the function, it often n...

AI Technical Summary

Problems solved by technology

[0005] In order to overcome the strict requirements on the function properties of classical optimization methods and the complex and difficult-to-solve local extremum problems existing in other hardw

Images