Natural Computing Method of Multispace Coevolution
SUN Xiao-qing1,JI Wei-dong1,LIN Ping2,XU Hao-tian1
1(Department of Computer Science and Information Engineering,Harbin Normal University,Harbin 150025,China)2(Harbin Medical Sciences University,Harbin 150086,China)
Abstract：In the natural calculation method,the population size is large and the calculation complexity is high;the population size is small and easy to fall into local optimum.In this paper,a natural computing method of multispace coevolution(MSC)is proposed.This method is suitable for all kinds of optimization algorithms based on population evolution,does not depend on the specific steps of algorithm evolution,and has universality.On the basis of the traditional evolution of biological population,the large population is decomposed into a limited number of small populations,some of which constitute the evolution space,and the other one constitutes the guidance space.The two spaces have different functions.The guidance space transmits the general information of experience to the evolution space through a specific way of information transmission,so that the whole population can co evolve.This strategy is applied to particle swarm optimization(PSO)and genetic algorithm(GA)respectively,and compared with standard particle swarm optimization(PSO),genetic algorithm(GA)and the current mainstream seven algorithms for large-scale optimization.In the high-dimensional function,the results show that the new evolutionary algorithm for population optimization has increased by 80% and universality compared with other algorithms.
2022, Vol. 43 
