0% found this document useful (0 votes)
16 views15 pages

Chapter 4

This document discusses methods for solving systems of nonlinear equations. It begins by defining a system of n nonlinear equations with n unknowns. The Jacobian matrix, which contains the partial derivatives of each equation with respect to each unknown, is introduced. An iterative method is then presented where the nonlinear system is linearized at each iteration using a Taylor series expansion. This results in a system of linear equations that can be solved for the unknowns at the next iteration. The iterations continue until the solutions converge within a specified tolerance. In summary, it presents an iterative method for solving nonlinear systems by linearizing and inverting the Jacobian matrix at each step.

Uploaded by

Abel Taye
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
0% found this document useful (0 votes)
16 views15 pages

Chapter 4

This document discusses methods for solving systems of nonlinear equations. It begins by defining a system of n nonlinear equations with n unknowns. The Jacobian matrix, which contains the partial derivatives of each equation with respect to each unknown, is introduced. An iterative method is then presented where the nonlinear system is linearized at each iteration using a Taylor series expansion. This results in a system of linear equations that can be solved for the unknowns at the next iteration. The iterations continue until the solutions converge within a specified tolerance. In summary, it presents an iterative method for solving nonlinear systems by linearizing and inverting the Jacobian matrix at each step.

Uploaded by

Abel Taye
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/ 15

Numerical Methods

(CENG-2084)
Chapter 4
Solutions of Systems of Nonlinear
Equations

3/23/2020 1
Systems
Systemsof
of Nonlinear Equations
Nonlinear Equations
•Locate the roots of a set of simultaneous nonlinear equations:

f1 ( x1 , x2 , x3 ,, xn )  0
f 2 ( x1 , x2 , x3 , , xn )  0

f n ( x1 , x2 , x3 ,, xn )  0
Example:
x12  x1 x2  10  f1 ( x1 , x2 )  x12  x1 x2  10  0
x2  3x1 x2 2  57  f 2 ( x1 , x2 )  x2  3x1 x2 2  57  0
3/23/2020 2
• First-order Taylor series expansion of a function with more than one variable:
f1(i ) f1( i )
f1(i 1)  f1( i )  ( x1( i 1)  x1( i ) )  ( x2( i 1)  x2( i ) )  0
x1 x2
f 2(i ) f 2(i )
f 2(i 1)  f 2(i )  ( x1( i 1)  x1( i ) )  ( x2( i 1)  x2( i ) )  0
x1 x2
• The root of the equation occurs at the value of x1 and x2 where f1(i+1) =0 and f2(i+1) =0
Rearrange to solve for x1(i+1) and x2(i+1)

f1(i ) f1(i ) f1(i ) f1(i )


x1(i 1)  x2(i 1)   f1(i )  xi  x2 ( i )
x1 x2 x1 x2
f 2(i ) f 2(i ) f 2(i ) f 2(i )
x1(i 1)  x2(i 1)   f 2(i )  xi  x2 ( i )
x1 x2 x1 x2

 f1(i ) f1(i )   f1(i ) f1( i ) 


   f1(i )   x1 
 x1 x2   x1( i 1)  x2   x1(i ) 
f 2 (i )   x2 (i 1) 
     f f 2(i )   x2( i ) 
 f 2( i )  2(i )   2(i )
f
 x x2   x x2 
 1  1
• Since x1(i), x2(i), f1(i), and f2(i) are all known at the ith iteration, this represents a set of
two linear equations with two unknowns, x1(i+1) and x2(i+1)
• You may use several techniques to solve these equations 3
General Representation for the solution of
Nonlinear Systems of Equations
 f1 f1 f1   Jacobian Matrix
f1 ( x1 , x2 , x3 ,, xn )  0  x 
x2 xn 
 1 
f 2 ( x1 , x2 , x3 , , xn )  0  f 2 f 2

f 2 
Solve this set of linear
J   x1 x2 xn 
      
equations at each iteration:
 f f n f n 
 n  
f n ( x1 , x2 , x3 ,, xn )  0  x1 x2 xn  [ J i ]{ X i 1}  {Fi }  [ J i ]{ X i }

 f1 f1 f1   f1 f1 f1 


 x   x   f ( x , x , x ,  , x )    x1 
xn     x 
1 1 1 2 3 n
x2   x2 xn  
 1   1 
 f 2 f 2 f 2      f f 2 f 2   
  x2   f 2 ( x1 , x2 , x3 ,, xn )   2   x2 
 x1 x2 xn           x1 x2 xn    
               
 f f n f n         f f n f n    
 n    xn   f n ( x1 , x2 , x3 , , xn )  n    xn 
 x1 x2 xn  i i 1 i  x1 x2 xn  i i

{X}(i+1) values are found by solving this system of linear equations.


These iterations will continue until the convergence is reached;
which will happen when the approximate relative error (defined on the vector of values) falls below the desired error tolerance εtol 4
3/23/2020
Solution of Nonlinear Systems of Equations

Solve this set of linear equations at each iteration: [ J i ]{ X i 1}  {Fi }  [ J i ]{ X i }

Rearrange: [ J i ]{ X i 1  X i }  {Fi }
[ J i ]{X i 1}  {Fi }

 f1 f1 f1 


 x    x1   f1 ( x1 , x2 , x3 ,  , xn ) 
x2 xn    
 1 
 f 2 f 2 f 2     
 x2   f 2 ( x1 , x2 , x3 , , xn ) 
 x1 x2 xn        
  
      
 f f n f n       
 n   xn   f n ( x1 , x2 , x3 ,  , xn )
 x1 x2 xn  i i 1 ii

3/23/2020 5
Notes on the solution of
Nonlinear Systems of Equations
n non-linear equations f1 ( x1 , x2 , x3 ,, xn )  0  f1
 x
f1

f1 
n unknowns to be solved: x2 xn 
 1 
f 2 ( x1 , x2 , x3 , , xn )  0  f 2 f 2

f 2 
Jacobian J   x1 x2 xn 
      
Matrix:  f f n f n 
f n ( x1 , x2 , x3 ,, xn )  0  n
 x1 x2
 
xn 

• Need to solve this set of Linear Equations

[ J i ]{X i 1}  {Fi }


in each and every iterative solution of the original Non-Linear system of equations.

• For n unknowns, the size of the Jacobian matrix is n2 and the time it takes to solve
the linear system of equations (one iteration of the non-linear system) is
proportional to O(n3)
• There is no guarantee for the convergence of the non-linear system. There may also
be slow convergence.
• In summary, finding the solution set for the Non-Linear system of equations is an
extremely compute-intensive task.
3/23/2020 6
Iterative method
Fixed -Point iteration for Functions of Several Variables
Procedures used for solving non- linear problem by iterative method
• the 𝑖 𝑡ℎ equation is solved for 𝑥𝑖, the system is changed into the fixed-point
problem
• Using the initial guess 𝑥2 , 𝑥3 … solve for 𝑥1
• Using the values of 𝑥1 from step ii and 𝑥3 , 𝑥4, … solve for 𝑥2
• Using the value of x1 from step ii and that of x2 from step iii solve for x3
• Repeat this procedure until the required accuracy is achieved. /interval is
smaller than some specified tolerance/
Example : solve the system of nonlinear equation by iterative method
1
3𝑥1 − cos 𝑥2 𝑥3 − = 0
2
2 2
𝑥1 − 81 𝑥2 + 0.1 + 𝑠𝑖𝑛𝑥3 + 1.06 = 0
−𝑥1 𝑥2
10𝜋 − 3
𝑒 + 20𝑥3 + =0
3

3/23/2020 7
Solution
If the 𝑖𝑡ℎ equation is solved for 𝑥𝑖 ,
the system is changed into the
fixed-point problem
1 1
𝑥1 = cos 𝑥2 𝑥3 +
3 6
1
𝑥2 = 𝑥1 2 + 𝑠𝑖𝑛𝑥3 + 1.06
9
− 0.1
1 −𝑥 𝑥 10𝜋 − 3
𝑥3 = − 𝑒 1 2 −
20 60

3/23/2020 8
Solution
If the 𝑖𝑡ℎ equation is solved for 𝑥𝑖 , To approximate the fixed point𝑡p, we
choose 𝑥 0 = (0.1,0.1, −0.1) . The
the system is changed into the
sequence of vectors generated by
fixed-point problem (𝑘)
1 (𝑘−1) (𝑘−1)
1
1 1 𝑥1 = cos 𝑥2 𝑥3 + ,
𝑥1 = cos 𝑥2 𝑥3 + 3 6
3 6 1
1 𝑥2 (𝑘) = (𝑥1 (𝑘−1) )2 + sin 𝑥3 (𝑘−1) + 1.06
𝑥2 = 𝑥1 2 + 𝑠𝑖𝑛𝑥3 + 1.06 − 0.1 9
9 − 0.1,
1 −𝑥 𝑥 10𝜋 − 3 1 −𝑥 (𝑘−1) 𝑥 (𝑘−1) 10𝜋 − 3
𝑥3 = − 𝑒 1 2 − 𝑥3 (𝑘) =− 𝑒 1 2 −
20 60 20 60

3/23/2020 9
Solution
If the 𝑖𝑡ℎ equation is solved for 𝑥𝑖 , To approximate the fixed point𝑡p, we
choose 𝑥 0 = (0.1,0.1, −0.1) . The
the system is changed into the
sequence of vectors generated by
fixed-point problem (𝑘)
1 (𝑘−1) (𝑘−1)
1
1 1 𝑥1 = cos 𝑥2 𝑥3 + ,
𝑥1 = cos 𝑥2 𝑥3 + 3 6
3 6 1
1 𝑥2 (𝑘) = (𝑥1 (𝑘−1) )2 + sin 𝑥3 (𝑘−1) + 1.06
𝑥2 = 𝑥1 2 + 𝑠𝑖𝑛𝑥3 + 1.06 − 0.1 9
9 − 0.1,
1 −𝑥 𝑥 10𝜋 − 3 1 −𝑥 (𝑘−1) 𝑥 (𝑘−1) 10𝜋 − 3
𝑥3 = − 𝑒 1 2 − 𝑥3 (𝑘) =− 𝑒 1 2 −
20 60 20 60

𝒌 𝑿𝟏 (𝒌) 𝑿𝟐 (𝒌) 𝑿𝟑 (𝒌) /𝑿𝟏 (𝒌) − 𝑿𝟏 (𝒌−𝟏) / /𝑿𝟐 (𝒌) − 𝑿𝟐 (𝒌−𝟏) / /𝑿𝟑 (𝒌) − 𝑿𝟑 (𝒌−𝟏) /
0 0.1 0.1 -0.1

1 0.49998333 0.00944115 -0.52310127 0.399983333 0.09055885 0.4231

2 0.49999593 2.5568E-05 -0.52336331 1.26014E-05 0.009415582 0.00026

3 0.50000000 1.2337E-05 -0.52359814 4.06505E-06 1.3231E-05 0.00023

4 0.50000000 3.4168E-08 -0.52359847 2.28887E-11 1.23026E-05 3.3E-07


3/23/2020 10
5 0.50000000 1.6487E-08 -0.52359877 6.9541E-12 1.76809E-08 3.1E-07
One way to accelerate convergence of the fixed-point iteration is to use the latest estimates x1
(k), … , xi-1 (k) instead of x1(k-1), … , xi-1(k-1) to compute xi(k), as in the Gauss-Seidel
method for linear systems.
The component equations then become
(𝑘)
1 (𝑘−1) (𝑘−1)
1
𝑥1 = 𝑐𝑜𝑠 𝑥2 𝑥3 +
3 6
(𝑘)
1
𝑥2 = (𝑥1 (𝑘−1) )2 + 𝑠𝑖𝑛 𝑥3 (𝑘−1) + 1.06 − 0.1
9
1 −𝑥 (𝑘−1)𝑥 (𝑘−1) 10𝜋 − 3
𝑥3 (𝑘) = − 𝑒 1 2 −
20 60
So the convergence was indeed accelerated for this problem by using the Gauss-Seidel
method. However, this method does not always accelerate the convergence.

𝒌 𝑿𝟏 (𝒌) 𝑿𝟐 (𝒌) 𝑿𝟑 (𝒌) /𝐗 (𝒌) −𝐗 (𝒌−𝟏) /


0 0.1000000 0.1000000 -0.1000000
1 0.49998333 0.02222979 -0.52304613 0.423046
2 0.49997747 2.8154E-05 -0.52359807 0.022202
3 0.50000000 3.7622E-08 -0.52359877 2.81E-05
4
3/23/2020 0.50000000 5.028E-11 -0.52359878 3.76E-08 11
Example:- Solve the nonlinear system by Newton’s method
1
3x1 − cos x2 x3 − =0
2
x1 2 − 81 x2 + 0.1 2 + sinx3 + 1.06 = 0
10π − 3
e−x1 x2 + 20x3 + =0
3
Newton’s method will be used to obtain this approximation using the initial
approximation is 𝑥 (0) = 0.1,0.1, −0.1 𝑎𝑛𝑑
𝑭(𝒙𝟏 , 𝒙𝟐 , 𝒙𝟑 ) = (𝑓1 (𝑥1 , 𝑥2 , 𝑥3 ), 𝑓2 (𝑥1 , 𝑥2 , 𝑥3 ), 𝑓3 (𝑥1 , 𝑥2 , 𝑥3 ))𝑡 ,
Where
1
𝑓1 𝑥1 , 𝑥2 , 𝑥3 = 3x1 − cos x2 x3 −
2
2 2
𝑓2 𝑥1 , 𝑥2 , 𝑥3 = x1 − 81 x2 + 0.1 + sinx3 + 1.06
−x1 x2
10π − 3
𝑓3 𝑥1 , 𝑥2 , 𝑥3 = e + 20x3 +
3
And
The Jacobian matrix 𝑱(𝒙) for this system
3 𝑥3 sin𝑥2 𝑥3 𝑥2 sin𝑥2 𝑥3
𝐽 𝑥1 , 𝑥2 , 𝑥3 = 2𝑥1 −162(𝑥2 + 0.1) cos𝑥3
−𝑥2 e−x1x2 −𝑥1 e−x1x2 20

3/23/2020 12
 Thus, at the 𝑘𝑡ℎ step, the linear system 𝑱(𝐱 𝐤−𝟏 )𝐲 𝐤−𝟏 = −𝐅(𝐱 𝐤−𝟏 )
𝒚𝟏 𝒌−𝟏
𝒌−𝟏 = −(𝑱(𝒙𝟏 𝒌−𝟏 , 𝒙𝟐 𝒌−𝟏 , 𝒙𝟑 𝒌−𝟏 ))−𝟏 𝑭(𝒙𝟏 𝒌−𝟏 , 𝒙𝟐 𝒌−𝟏 , 𝒙𝟑 𝒌−𝟏 )
𝒚𝟐
𝒚𝟑 𝒌−𝟏
1
𝑥3 sin𝑥2 𝑥3 𝑦1 3x1 − cos x2 x3 −
3 𝑥2 sin𝑥2 𝑥3 2
2𝑥1 −162(𝑥2 + 0.1) cos𝑥3 𝑦2 2 2
= − x1 − 81 x2 + 0.1 + sinx3 + 1.06
−x1 x2 𝑦3
−𝑥2 e −𝑥1 e−x1x2 20 −x1 x2
10π − 3
e + 20x3 +
3
Taking 𝑥1 (0) = 0.1, 𝑥2 (0) = 0.1 𝑎𝑛𝑑 𝑥3 (0) = −0.1 as initial approximation, the above
equation reduces to
3 9.99𝑥10−4 −9.99𝑥10−4 𝑦1 1.9995
0.2 −32.4 0.995 ∗ 𝑦2 = 2.2698
−0.099 −0.099 20 𝑦3 −8.46
Solving this linear system of equation, results in
𝑦1 = 0.39987, 𝑦2 = − 0.080533 𝑎𝑛𝑑 𝑦3 = −0.42121.

3/23/2020 13
Use this results, the next approximation for the unknown variables can be calculated as
𝒙𝟏 𝟏 𝒙𝟏 𝟎 𝒚𝟏 𝟎
𝒙𝟐 𝟏 = 𝒙𝟐 𝟎 + 𝒚𝟐 (𝟎)
𝟏 𝟎 𝒚𝟑 𝟎
𝒙𝟑 𝒙𝟑

𝒙𝟏 𝟏 = 𝒙𝟏 𝟎 + 𝒚𝟏 𝟎 = 𝟎. 𝟏 + 𝟎. 𝟑𝟗𝟗𝟖𝟕 = 𝟎. 𝟒𝟗𝟗𝟖𝟕
𝒙𝟐 𝟏 = 𝒙𝟐 𝟎 + 𝒚𝟐 𝟎 = 𝟎. 𝟏 − 𝟎. 𝟎𝟖𝟎𝟓𝟑𝟑𝟏 = 𝟎. 𝟎𝟏𝟗𝟒𝟔𝟔
𝒙𝟑 𝟏 = 𝒙𝟑 𝟎 + 𝒚𝟑 𝟎 = −𝟎. 𝟏 − 𝟎. 𝟒𝟐𝟏𝟐𝟏 = −𝟎. 𝟓𝟐𝟏𝟓𝟐𝟎

𝒌 𝑿𝟏 (𝒌) 𝑿𝟐 (𝒌) 𝑿𝟑 (𝒌) /𝐗 (𝒌) −𝐗 (𝒌−𝟏) /


0 0.1000000 0.10000000 -0.10000000
1 0.4998700 0.02222979 -0.52304613 0.423046
2 0.50004593 2.8154E-05 -0.52359807 0.022202
3 0.50000034 3.7622E-08 -0.52359877 1.54x10-5
4 0.50000000 0.00000000 -0.52359878 1.24-08
5 0.50000000 0.00000000 -0.52359878 0.0000000

The example illustrates that Newton's method can converge very rapidly once an
approximation is obtained that is near the true solution.
3/23/2020 14
Thank You

3/23/2020 15

You might also like