Based On The Extended Particle Swarm Algorithm Population Structure Has To Self-organization

Extended Particle Swarm Optimization (EPSO) algorithm is a newalgorithm by simulating intelligent behavior of biological community,asbiological community is a complicated and self-organizing network structure,ifEPSO algorithm can simulate self-organizing evolutionary process of biologicalcommunity,more complex behaviors will emerge and it will really simulate theintelligence behavior of biological community.For the shortage of undirectedtopological structure in EPSO algorithm,directed self-organizing topologicalstructure are constructed in this paper,and the topological structurecharacteristics and the relationships between evolutions of characteristicmeasures of directed topological structure and performance of EPSO algorithmare studied.Firstly,in order to construct directed self-organizing topologicalstructures,various static directed topological structures were constructed forresearch influence on the performance of the EPSO algorithm fromcharacteristic measures of static directed topological structures.The conclusionthat the outdegree and the fitness values of particles are the important factorsinfluence on the local search and global search ability of EPSO algorithm wasdrawn by simulation experiment.Secondly,according to the conclusion of staticdirected topological structures in EPSO algorithm,EPSO-DSOTIS algorithm isproposed,in the algorithm,the evolution mechanism of directed self-organizingtopological structures under invariable size that the indegree of particles selectconnection object according to the fitness values and the outdegree of particlesgradually decreasing were established.Simulation results show that theevolutionary mechanisms improved the local search ability of EPSO algorithmby improved the speed of the optimal fitness information communication.Finally,in order to further improve the local search ability of EPSO algorithm,asthe dynamic complex networks exist in the actual social life and the thought ofsurvival of the fittest of biological group,EPSO-DSOTDC algorithm on the basisof EPSO-DSOTIS algorithm is proposed,in EPSO-DSOTDC algorithm the evolution mechanism of directed self-organizing topological structures based onthe dynamic changes of node deletion and compensation were established.Theexperimental results show that with the algorithm evolution constantly,thediversity values of the topological structure decrease gradually,it proved that theEPSO-DSOTDC algorithm are further improved the local search ability of theEPSO algorithm.