Scribd
Upload
12 views13 pages

#11 Modified Newton Method

Uploaded by

Robert Dalton
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
12 views13 pages

#11 Modified Newton Method

Uploaded by

Robert Dalton
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 13

ME 782 Design Optimization

Modified Newton’s Method


Modified Newton’s Method
xi +1 = xi − H i−1f (xi )
Search direction using the Newton’s method

xi +1 = xi −  H f (xi ) −1
i

Search direction modified by using an optimum 


using the modified Newton’s method
df  xi +1 (  ) 
=0
d

8/18/2024 D.N. Manik 2


Modified Newton’s Method
Start with an initial point 𝐱0
x1 = x 0 +  1s1
s1 = −H ( x 0 ) f ( x 0 )

df  x1 (  ) 
=0
d
x1 = x 0 +  1s1
Repeat until convergence
8/18/2024 D.N. Manik 3
Newton’s Method
Consider the following objective function

f ( x1 , x2 ) = x12 + 2 x22 − 4 x1 − 2 x1 x2

 2 x − 2 x2 − 4 
f ( x1 , x2 ) =  1 
 4 x2 − 2 x1 

 2 −2 
H= 
 −2 4 
Beginning from the starting point (1,1), carry out three

iterations using the modified Newton’s method. Wherever

it is necessary to determine  , assume it as 0.4.


8/18/2024 D.N. Manik 4
Newton’s Method
1
x0 =  
1
 −4 
f ( x 0 ) =  
2
 2 −2 
H ( x0 ) =  
 − 2 4 
x1 = x 0 −   H ( x 0 )  f ( x 0 )=
−1

−1
1  2 −2   −4   2.2 
1 − 0.4  −2 4   2  = 1.4 
      
8/18/2024 D.N. Manik 5
Newton’s Method
 −2.4 
f ( x1 ) =  
 1.2 
 2 −2 
H ( x1 ) =  
 −2 4 
x 2 = x1 −   H ( x 0 )  f ( x1 ) =
−1

−1
 2.2   2 −2   −2.4   2.92 
1.4  − 0.4  −2 4   1.2  = 1.64 
       
8/18/2024 D.N. Manik 6
Newton’s Method
 −1.44 
f ( x 2 ) =  
 0.72 
 2 −2 
H ( x2 ) =  
 −2 4 
x3 = x 2 −   H ( x 2 )  f ( x 2 )=
−1

−1
 2.92   2 −2   −1.44   3.352 
1.64  − 0.4  −2 4   0.72  = 1.7840 
       
8/18/2024 D.N. Manik 7
Modified Newton’s Method
Minimize the Himmelblau function
f ( x1 , x2 ) = x + x + 2 x1 x2 ( x1 + x2 ) − 21x
4
1
4
2
2
1

−13 x − 14 x1 − 22 x2 + 170
2
2 0 
x0 =  
 4 x13 + 2 x22 + 4 x1 x2 − 42 x1 − 14 
0 
f =  3 
4
 2x + 2 x1
2
+ 4 x x
1 2 − 26 x2 − 22 

12 x12 + 4 x2 − 42 4 ( x1 + x2 ) 
H= 
 4 ( x1 + x2 ) 12 x2 + 4 x1 − 26 
2

8/18/2024 D.N. Manik 8


Modified Newton’s Method
0   −14  1
x0 =   f ( x 0 ) =  
0   −22 
 −42 0 
H ( x0 ) =  
 0 −26 
 −0.3665
s1 = −  H ( x 0 )  f ( x 0 ) = 
−1

−0.9304 
 

 −1.0982 
x1 = x 0 +  1s1 =    1 = 2.9961
 −2.7876 
8/18/2024 D.N. Manik 9
Modified Newton’s Method
 −1.0982   54.6120  2
x1 =   f ( x1 ) =  
 −2.7876   −21.5138
 −38.6792 −15.5431
H ( x1 ) =  
 −15.5431 62.8576 
−1 0.8790 
s 2 = −  H ( x1 )  f ( x1 ) =  
 0.4769 

 3.4115   2 = 5.1305
x 2 = x1 +  2s 2 =  
 −0.3410 
8/18/2024 D.N. Manik 10
Modified Newton’s Method
 3.4115   −2.8919  3
x2 =   f ( x 2 ) =  
 −0.3410   5.3304 
96.2920 12.2817 
H ( x2 ) =  
12.2817 −10.9587 

−1  −0.0614 
s3 = −  H ( x 2 )  f ( x 2 ) =  
 0.9981 
3.2759 
x3 = x 2 +  3s3 =    3 = 2.2065
1.8613 
8/18/2024 D.N. Manik 11
Modified Newton’s Method
3.2759   20.3535 4
x3 =   f ( x 3 ) =  
1.8613   1.2526 
94.2244 20.5488 
H ( x3 ) =  
 20.5488 28.6769 
 −0.8806 
s 4 = −  H ( x3 )  f ( x3 ) = 
−1

 0.4738 
 2.9998
x 4 = x3 +  4s 4 =    4 = 0.3136
 2.0099 
8/18/2024 D.N. Manik 12
Conclusion

8/18/2024 D.N. Manik 13

You might also like