Research On Improvement And Application Of Particle Swarm Optimization

Particle Swarm Optimization(PSO)as a highly-efficient stochastic swarm intelligence optimization method has been extensively applied in various fields of scientific research and engineering practice.In this paper,an improved PSO algorithm based on the switch between the convergent operator and the divergent operator has been studied against the difficulty in controlling the alternated occurrence of convergent state and divergent state in the iterative process of PSO in accordance with convergent analysis results of time invariant transfer matrix,so as to balance the "exploration" capacity and the "development" capacity of the swarm.The contrast experiment of simulating optimization shows that the controllability of the convergent state and the divergent state and the efficiency of global convergence can be found in the improved PSO algorithm.And the method of PSO application in target positioning and target as a critical problem in the field of computer vision has been also studied.Target positioning is to locate the target of interest in the image,while target tracking is to track the moving target of interest in the video in real time and determine its moving trail.Research on target positioning and tracking methods has a wide range of application prospects.For instance,target positioning and tracking methods are applied in video intelligent real-time monitoring systems,tracking and focusing of moving targets in cameras,drone positioning and tracking,positioning of victims in rescue management systems,and acquisition of deploy data of enemy's military presence using video information in military matters.Moreover,an improved target function has been studied against the identification of a single target,which can be used for identifying irregular objects.Meanwhile,a particle swarm optimization algorithm based on multi-swarm cooperative competition has been also studied to raise the efficiency of the PSO algorithm through dynamically assigning the number of particles in each group.In addition,a multi-swarm adaptive method that can track multiple adjacent targets has been studied against the problem of tracking a plurality of targets.When multiple targets are far apart,each swarm only tracks a single target;and when two tracked targets are close,one swarm is responsible for tracking the center of the two targets,while the other is for preventing peripheral targets from escaping the coverageof particles.The effectiveness of multi-swarm cooperative particle swarm optimization algorithm has proven in the simulating experiment of locating and tracking multiple irregular ants.