Numerical Analysis: Difference between revisions

From David's Wiki
(Created page with "Numerical Analysis ==Nonlinear Equations== ===Continuation/Homotopy Methods=== Reference Numerical Optimization by Nocedal and Wright (2006)<br> Also known as zero-path follo...")
 
No edit summary
Line 5: Line 5:
Reference Numerical Optimization by Nocedal and Wright (2006)<br>
Reference Numerical Optimization by Nocedal and Wright (2006)<br>
Also known as zero-path following.<br>
Also known as zero-path following.<br>
If you want to solve <math>r(x)=0</math> when <math>r(x)</math> is difficult (i.e. has non-singular Jacobian) then you can use this method.
If you want to solve <math>r(x)=0</math> when <math>r(x)</math> is difficult (i.e. has non-singular Jacobian) then you can use this method.<br>
Define the Homotopy map <math>H(x, \lambda) = \lambda r (x) + (1 \lambda)(x a)</math>
Define the Homotopy map<br>
<math>H(x, \lambda)=\lambda r(x) + (1-\lambda)(x-a)</math>

Revision as of 13:51, 19 October 2019

Numerical Analysis

Nonlinear Equations

Continuation/Homotopy Methods

Reference Numerical Optimization by Nocedal and Wright (2006)
Also known as zero-path following.
If you want to solve \(\displaystyle r(x)=0\) when \(\displaystyle r(x)\) is difficult (i.e. has non-singular Jacobian) then you can use this method.
Define the Homotopy map
\(\displaystyle H(x, \lambda)=\lambda r(x) + (1-\lambda)(x-a)\)