ABSTRACT
The essence of PSO algorithm easily falling into local optimum is the existence of particle “overconcentrate” phenomenon in late stage. And the probability of relying on existing pattern to break local optima is small. Therefore, do artificial migration through extreme perturbation operator to strengthen the global search ability of particles [5]. That is:
P r Pid id3 (10)
P r Pgd gd4 (11)
r t T
U t T3 0 0
1 =
>
⎧⎨⎩ ,
( ,0 ), (12)
r t T
U t T g g
1 =
>
⎧ ⎨⎪ ⎩⎪
,
( ,0 ),
(13)
where Pid = individual extreme; Pgd = global extreme; t0, tg are individual extreme and global extreme evolutionary stagnation step number; T0, Tg = the need of disturbance stagnation step number threshold for individual extreme and global extreme. U(0,1) = uniform random function.
