Abstract: A certain number particle swarm is subdivided into three equal groups，then they evolve with different rules，
which are particle swarm optimizer，a modified particle swam optimizer and a cloud adaptive particle swarm optimizer.It not
only keeps the independence of the particle swarm and the superiority of the optimizer，but also not increases the complexity
of algorithm.Meanwhile，the super-society parts are put forward and then the velocity formula is redefined.At the same
time，variation expansion methods and disturbance operation are introduced.The experimental results show that the new algorithm
has advantages of convergence property，convergence speed，accuracy and the stability of convergence effective.