Small white before use Matlab to write a code of particle swarm optimization (pso), running is good, but according to the previous ideas to Python code running effect is very poor, also found in the process of iteration func1 (self. P [I]) and self p_best [I] and func1 (self) g) and self g_best values corresponding to the up, which can have a great god to help us have a look?

 
The import numpy as np 
Import the random 


Def func1 (z) : 
Total=0 
For I in range (0, z.s considering) : 
[I] total=z * * 2 + total 
Return the total 


The class PSO: 
Def __init__ (self, the size, dim, iter_num, C1 and C2, W, X_max, X_min, V_max, V_min) : 
Self. Size=size 
Self. Dim=dim 
Self. Iter_num=iter_num 
The self. The C1=C1 
The self, C2=C2 
Self. W=W 
Self. X_max=X_max 
Self. X_min=X_min 
Self. V_max=V_max 
Self. V_min=V_min 

Self. X=np. Random. The random ((self. The size, the self. Dim)) * (X_max - X_min) + X_min 
Self. V=np. Random. The random ((self. The size, the self. Dim)) * (V_max - V_min) + V_min 
Self. P=self. X 
Self. P_best=[func1 (self. X [I]) for I in range (0, the self. The size)] 
Self. G_best=self. P_best [0] + 1 
For I in range (0, the self. The size) : 
If self. P_best [I] Self. G=self. [I] p 
Self. G_best=self. P_best [I] 
Self. Gb=[] 
The self. The fitness=0 

Def update (self) : 
For I in range (self. Iter_num) : 
For j in range (self. The size) : 
The self. The fitness=func1 (self. X [j]) 

# update individual optimal location, optimal value 
If the self. The fitness & lt; Self. P_best [j] : 
Self. P=self [j]. [j] x 
Self. P_best [j]=self. Fitness 

# update group, the optimal location, optimal value 
If self. P_best [j] Self. G=self. P [j] 
Self. G_best=self. P_best [j] 

# update speed, position 
Self. V=self [j]. W * self in v + self [j]. The C1 * random in the random () * (self. P [j] - self. X [j]) \ 
+ self. C2 * random. The random () * (self. G - self. X [j]) 
Self. X [j]=self. V + self [j]. X [j] 


 # boundary conditionsFor k in range (self. Dim) : 
If self. V [j] [k] & gt; Self. V_max or self. V [j] [k] Self. V [j] [k]=random. The random () * (self. V_max - self. V_min) + self. V_min 
If self. X [j] [k] & gt; Self. X_max or self. X [j] [k] Self. X [j] [k]=random. The random () * (self. X_max - self. X_min) + self. X_min 

Self. Gb. Append (self g_best) 

Return the self. J g, self. G_best 

 
The from PSO import PSO 

Size=10 
Dim=3 
Iter_num=200 
C1=1.5 
C2=1.5 
W=0.8 
X_max=20 
X_min=- 20 
V_max=10 
V_min=- 10 

Pso=pso (size, dim, iter_num, C1 and C2, W, X_max, X_min, V_max, V_min) 
Best_pos, best_fitness=pso. The update () 

In addition, I found a problem, why this way only changed the value of the a and b values has changed?



And this kind of situation, both values are not binding



Does anyone know?