A Novel Transformers Fault Diagnosis Method Based on Probabilistic Neural Network and Bio-Inspired Optimizer

. 2021 May 23;21(11):3623. doi: 10.3390/s21113623

Algorithm 1 Improved salp swarm algorithm.

1:
Initialization parameters: population size N, dimension d, maximum number of iterations T.
2:
Generate the initial population X by Equation (1);
3:
Calculate the fitness value for each individual;
4:
while $t < = T$ do
5:
Update $c_{1}$ by Equation (3) and $r_{1}$ by Equation (8);
6:
for $i = 1 : n$ do
7:
if $X i (l e a d e r)$ then
8:
Update random numbers $c_{2}$ and $c_{3}$ ;
9:
Update the position of the leader salp as in Equation (2);
10:
else
11:
Update random numbers $r_{2}$ , $r_{3}$ and $r_{4}$ ;
12:
Update the position of the follower salp as in Equation (7);
13:
end if
14:
Calculation $D_{o p}$ using Equation (9);
15:
end for
16:
if $δ_{0} > 0.5$ then
17:
$X = X \cdot D_{o p}$ ;
18:
end if
19:
Set $t = t + 1$ ;
20:
end while

Output: Best classification and predication results.