Self-adaptive genetic particle swarm hybrid algorithm optimization method

Abstract

The invention provides a self-adaptive genetic particle swarm hybrid algorithm optimization method. The self-adaptive genetic particle swarm hybrid algorithm optimization method includes: calculatingthe density and the radius of a center region of a parent population in a genetic algorithm, and distinguishing whether the parent population is in the overall centralized distribution, the local centralized distribution or the uniform distribution; performing a selection operation of the genetic algorithm, and selecting a parent individual to be evolved; establishing computational formulas of thecrossover probability and the mutation probability according to the three distributions of the parent population; performing crossover and mutation operations according to the established crossover and mutation probability formulas so as to achieve chromosome recombination and gene mutation, and forming an offspring individual; selecting a part of individuals with high fitness from a part of offspring individuals to perform the particle swarm algorithm to form offspring particles, and combining the offspring individuals and the offspring particles into an offspring population and saving the optimal individual thereof. The invention adaptively adjusts crossover probability mutation probability parameter values in the genetic particle swarm hybrid algorithm, so that the convergence speed and the convergence precision are greatly improved.

Description

technical field [0001] The invention relates to the technical field of computing, in particular to an optimization method of an adaptive genetic particle swarm hybrid algorithm. Background technique [0002] With the development of optimization theory, intelligent optimization algorithms represented by genetic algorithm and particle swarm optimization have been developed rapidly. Genetic algorithm and particle swarm optimization, as two widely used intelligent optimization methods, have their own advantages and disadvantages. Genetic algorithm has a wide range of space search capabilities, but the convergence speed is slow; particle swarm optimization has the advantages of easy implementation and fast convergence speed, but it is easy to fall into the local optimal solution. Therefore, in order to make full use of the advantages of these two algorithms, many scholars have proposed a hybrid genetic particle swarm optimization algorithm that combines genetic algorithm and par...

AI Technical Summary

Problems solved by technology

However, at present, the two parameters of crossover probability and mutation probability in the genetic particle swarm optimization algorithm mainly depend on the experience of the operator to take fixed values, and the rationality and scientificity of the values ​​are lack of effective research, which is to a certain extent It affects the accuracy and speed of the algorithm, and has become the most important factor restricting the development of the genetic particle swarm optimization algorithm.

Images