Downloaded From: www.EasyEngineering.

net

ww
w.E
a syE
ngi
nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

ww
w.E Input

Input

a syE
Input layer
Input

Output layer

ngi Input
Hidden layer

nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

Axon hillock

Soma Axon

Dendrite Nucleus Terminal buttons

ww
w.E
a syE
ngi
nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

ww
w.E
a syE
ngi
nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

Neural networks

Feed-forward networks Recurrent networks

ww
Single layer
perception w.E
Multilayer
perception
Radial basis
function net
Competitive
networks
Kohonen
SOM
ART models

a syE
ngi
nee
rin
g.n
et
N
Ii Wi b
i 1
 Downloaded From: www.EasyEngineering.net

0
Bias 1 W
1
1 W
X
2
W Activation function
2 (more on this later)
X

3  Output
Inputs X

3
W

ww X
m m
W

w.E
a syE
ngi
nee
rin
xi g.n
et
n
w i xi
i 1

n
w i xi w0 ,w0
i 1

n
w i xi , x0 1
i 1

wi xi
w0
 Downloaded From: www.EasyEngineering.net

X0 = 1
X1 W1 W0 = – 
W2 X1 W1
X2
W3 0 W2
X2
0
X3 
X3 W3 

w0

ww 1 if x 0,
1 otherwise

w.E 1 if x 0,
0 otherwise

a syE
ngi
nee
rin
g.n
wi
wi t i Xi
et
ti
 Downloaded From: www.EasyEngineering.net

ww
w.E
a syE
ngi
w1 w2 w0 w0
nee
w1
w1
w2
w2
w0
w0
w0
w0
w2
w1 rin
w1 w2 w0 w0 w1 w2 g.n
et

1
g(x)
1 e x
 Downloaded From: www.EasyEngineering.net

XOR

e.g. OR AND

ww
w.E
a syE
ngi
nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

ww
2

1
2

1
2

0
w.E 0 0

–1

–2
–10 0
a 10
–1

syE
–2
–10 0 10
–1

–2
–10 0 10

Logistic function
ngi
Hyperbolic tangent
function
Identify function

nee
1 rin
1 e x
g.n
1 e
1 e
x
x
et
 Downloaded From: www.EasyEngineering.net

x1 W1j
W2j Sum | Threshold
x2 Activation xj
xj

ww x3 W
3j

w.E
Xj

xj
a
i

f(X j )
W ij
1
Wj

syE
1 exp ( X j )
ngi
xi nee
W ij
Wj rin
Input signals

g.n
W ij
X1
et Y1

X2 Y2

Xm Yp

Input Output
layer layer
Hidden
layer

Back-propagation of error correction

p th

Ep (d k xk) 2
k

dk xk
p th

ww i
Xi
Ep

w.E
E (t)
w ij (t)
a E(t)
a i (t) syE
a i (t)
w ij (t)

ngi
nee
rin
g.n
et
wi
xi

n
w i xi
i 1

w0
 Downloaded From: www.EasyEngineering.net

–1 +1 +1

W1
Level
W0
W2

+1 Output
W3  –1
–
Error Quantizer
Gains 
Input Summer +
pattern
switches
–1 +1

ww Reference
switch

w.E
a syE wi

ngi
nee
rin wi

g.n
p th et
Ep (t p op)2

tp
op

Ep wi

Ep
(t p o p ) xi
wi

Ep wi p th

p wi (t p o p ) xi
 Downloaded From: www.EasyEngineering.net

ww
w.E
a syE
ngi
nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

ww
w.E W ij Xi

a syE
ngi
Input 1
nee
Weight
W1
rin
Weight g.n
Input 2 W2

Weight Sigmoid
et
Output

WN

Input N
Threshold 

2
f W i Xi b
i 1
 Downloaded From: www.EasyEngineering.net

1 if s 0
1 if s 0

ww
W 1X 1
W 1X 1
W 2 X2
W 2 X2
... W n X n
... W n X n

w.E
a syE
ngi
W d x nee
rin
g.n
W W W
et

Wi Wi d (n) X i (n)
 Downloaded From: www.EasyEngineering.net

ww
w.E XOR

a syE e.g. OR AND

ngi
nee
rin
g.n
(X 1 , X 2 )
X1 et
o o
o o
x o
x o o o
x o
x x o
X2
(X 1 , X 2 ) x x x
x o
L
 Downloaded From: www.EasyEngineering.net

x 1 , x 2 , ... x n

ww
w.E
a syE
ngi
nee
rin
g.n
et
O X

O O

X : Class I (y = 1)
O : Class II (y = –1)
 Downloaded From: www.EasyEngineering.net

ww
w.E X X

a syE O X

ngi
X : Class I (y = 1)
O : Class II (y = –1)

nee
rin
g.n
et

X O

O X

X : Class I (y = 1)
O : Class II (y = –1)
 Downloaded From: www.EasyEngineering.net

b w1 w2 0 (1)
b w1 w2 0 (2)
b w1 w2 0 (3)
b w1 w2 0 (4)

ww
w.E
a syE
ngi
nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

ww
w.E
a syE
ngi
nee
rin
g.n
et

Xk 1 xk
 Downloaded From: www.EasyEngineering.net

x
x3 4
x2

ww x1

w.Ex0

a syE
ngi
nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

ww
En E n ( w ji ) – E n ( w ji )
O( )
w ji
w.E
En
aE n ( w ji
syE
) – E n ( w ji – )
w ji 2
ngi O( 2
)

O(W ) 2
nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

ww
w.E
a syE
ngi
nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

ww
w.E
a syE
ngi
nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

x li xi x ui ,

x li x ui xi

ww
w.E
a syE
ngi
nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

Describe
problem

Generate
initial
solutions

ww
Step 1
Test: is initial
solution good enough?
Yes
Stop

w.E
step 2
a syE
No

Select parents
to reproduce

ngi
Step 3 Apply crossover process
nee
and create a set of offspring
Step 4
Step 5 Apply random mutation
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

ww
w.E x3 x2 5

a syE
ngi
nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

ww
w.E
a syE
ngi
Fi i i nee i

Fi i i i
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

ww
w.E
a syE
ngi
nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

3n

ww
M3 n

w.E
a syE M3 n
ngi
nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

ww
w.E
a syE
(S, t 1) (S)
ngi
f(S)
nee
(S, t 1) (S, t) n
fi rin
g.n
et
fi
f (s)
(S, t 1) (S, t)
favg

f avg

(favg c favg )
(S, t 1) (S, t)
favg

(S, t) (1 c)
 Downloaded From: www.EasyEngineering.net

(S, t) (S, t)(1 c) t

ww
w.E
a syE
ngi
nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

ww
w.E
a
1
+

2 syE +


2 +


9
1
ngi
2 – 4

nee
2 1

rin
g.n
et
 Downloaded From: www.EasyEngineering.net

ww
w.E
a syE
ngi
nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

ww
w.E
a syE
ngi
nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

ww
w.E
a syE
ngi
nee
rin
g.n
et
 Downloaded From: www.EasyEngineering.net

ww
w.E
a syE
ngi
nee
rin
g.n
et