1: \begin{abstract}

2: Recently, much progress

3:  has been made on particle swarm optimization (PSO). A number of works

4:  have been devoted to analyzing the convergence of the underlying algorithms.

5: Nevertheless, in most cases,

6: rather simplified hypotheses

7: are used. For example, it often assumes that the swarm has only one

8: particle. In addition, more often than not, the variables and the

9: points of attraction are assumed to remain constant throughout the

10: optimization process. In reality, such assumptions are often

11: violated. Moreover, there are no rigorous rates of

12: convergence results available to date for the particle swarm, to the best of our knowledge. In this paper,

13: we consider a general form of PSO algorithms, and

14: analyze asymptotic properties of the algorithms using stochastic

15: approximation methods. We introduce four coefficients and rewrite

16: the PSO procedure as a stochastic approximation type iterative

17: algorithm. Then we analyze its convergence using weak convergence

18: method. It is proved that a suitably scaled sequence of swarms

19: converge to the solution of an ordinary differential equation. We

20: also establish certain stability results. Moreover, convergence

21: rates are ascertained by using weak convergence method. A centered

22: and scaled sequence of the estimation errors is shown to have a

23: diffusion limit.

24: \end{abstract}