Research On Improved Particle Swarm Optimization Based On Scale-Free Network

Particle Swarm Optimization(PSO)is an optimization algorithm in intelligent computing.Once it is proposed,it receives attention from various parties.This algorithm simulates the process of bird foraging and treats food as an optimal solution to the optimization problem.In this way,the PSO algorithm The process of finding the optimal solution is actually the process of finding food for the birds.Particle swarm optimization is a relatively new algorithm,with simple concepts,few control parameters,fast convergence,and strong convergence.It has a good practical effect in solving many problems to be optimized,and has a wide range of applications in many areas:constrained optimization,multi-objective optimization,data mining,and network community discovery.The PSO algorithm has good outstanding characteristics,but there are also certain drawbacks:the local search ability is poor,the population diversity is reduced,and it is easy to fall into the local minimum.In order to overcome the shortcomings of trapping into local optimum,the particle swarm optimization algorithm is improved by changing the perspective of information transfer between particles.On the basis of studying the topology structure of particle swarm optimization,the scale-free 'characteristics of complex networks are studied.Introduced into the particle swarm algorithm.The work done in this paper is as follows:A particle swarm optimization algorithm based on BA network is built,and compared with All,models,ring network models,and von Neumann network model particle swarm algorithm.The experimental results illustrate the particles based on BA network.The group algorithm has no advantage in simple single-peak function optimization capability,but has certain advantages in the optimization of complex unimodal functions and multi-peak functions.In order to further improve the algorithm,four evolutionary models of scale-free networks are used in the particle swarm optimization algorithm:(1)The fitness model changes the disadvantage that the older nodes in scale-free networks have higher connectivity,and constructs a fitness-based model.The particle swarm algorithm(FPSO);(2)The growth model constructs a particle swarm optimization algorithm(GPSO)based on the growth model by adding multiple particles at a time to form a scale-free network;(3)Local area world evolution model The network in the domain world is connected instead of the entire network,and a local particle swarm algorithm(LWPSO)based on the local world is constructed.(4)The dynamic evolution model not only includes nodes but also nodes in the evolution of the network.Deleted,a dynamic evolutionary particle swarm algorithm(DESPSO)was constructed,and the effect of adding edges and reducing edges on the performance of the algorithm was studied.The Benchmark standard test ffunction is used to verify the performance of the algorithm.From the simulation results,the performance of the particle swarm optimization algorithm based on the dynamic evolution model has certain advantages in optimizing the complex multi-peak function.