Page 203 - Decision Making Applications in Modern Power Systems

166 Decision Making Applications in Modern Power Systems

wji
j
i
Output
Inputs

i
Output layer
i is input neuran
Input layer
j is hidden neuron
W is weight
Hidden layer
FIGURE 6.6 Feed-forward neural network algorithm structure.

Hidden layer
W ah H 1 W hy
mn no
Input layer Output layer
A1 H 2 Y1

A2 H 3 Y2

H 4

Am Yo

Hn

FIGURE 6.7 Architecture of PNN. PNN, Probabilistic neural network.

Probabilistic density function is given by

C t
1 X j A 2 A to j
f t ðAÞ 5 exp 2 2 ð6:46Þ
C t 2λ
o51
Applying Eq. (6.46) to the output vector H, the hidden PNN layer becomes

0 1
ah 2
P
2 A m 2W
m mn
H n 5 exp @ A ð6:47Þ
2λ 2