Changes

1,751 bytes added ,  19:30, February 3, 2010
Line 100: Line 100:  
In single variable calculus, we found that extrema, that is, maximum and minimum values for a function, could only occur at points where the derivative was 0.  Dealing with multivariate functions, it should come as no surprise, then, that extrema can only occur when all of the partial derivatives are 0, or, to say the same thing, that the gradient is the zero length vector <math>\vec{0}</math>.
 
In single variable calculus, we found that extrema, that is, maximum and minimum values for a function, could only occur at points where the derivative was 0.  Dealing with multivariate functions, it should come as no surprise, then, that extrema can only occur when all of the partial derivatives are 0, or, to say the same thing, that the gradient is the zero length vector <math>\vec{0}</math>.
   −
As in the single variable case, having <math>\nabla f(x_0,y_0,z_0)=0</math> isn't sufficient to demonstrate that an extreme value of a function exists.
+
As in the single variable case, having <math>\nabla f(x_0,y_0,z_0)=0</math> isn't sufficient to demonstrate that an extreme value of a function exists.  It is necessary to rule out the possibility of an '''inflection point''', and, as in the single variable case, this is done with an examination of the second-order derivative. 
 +
 
 +
In the single variable case, we recall, a function <math>y=f(x) \ </math> has an extrema at a point <math>x=x_0 \ </math> if <math>f'(x_0)=0 \ </math> and <math>f''(x_0)\neq0</math>.  On the other hand, we have an inflection point if <math>f'(x_0)=f''(x_0)=0 \ </math>.  See the illustration at right.
 +
 
 +
It should come as no surprise then that we examine the second-order derivative of a multivariate function to determine when an extrema has taken place.  We examine one of the most important operators in higher mathematics now, the [[Laplacian]]: <math>\Delta = \nabla^2</math>.
 +
 
 +
In Cartesian coordinates, we will have
 +
 
 +
<math>\Delta f = \frac{\partial^2 f}{\partial x^2}+\frac{\partial^2 f}{\partial y^2}</math>
 +
 
 +
and in three dimensions, we will of course add on the term <math>\partial_{zz}f</math>.  In polar coordinates, the Laplacian is
 +
 
 +
<math>\Delta f = {1 \over r} {\partial \over \partial r} \left( r {\partial f \over \partial r} \right) + {1 \over r^2} {\partial^2 f \over \partial \theta^2} \ </math>
 +
 
 +
and in spherical coordinates,
 +
 
 +
<math>\Delta f = {1 \over r^2} {\partial \over \partial r} \left( r^2 {\partial f \over \partial r} \right) + {1 \over r^2 \sin \phi} {\partial \over \partial \phi}  \left( \sin \phi {\partial f \over \partial \phi} \right) + {1 \over r^2 \sin^2 \phi} {\partial^2 f \over \partial \theta^2}</math>
 +
 
 +
where we maintain our convention of using <math>\theta</math> for the azimuth and <math>\phi</math> for the zenith angle.  For more on how these different forms of the Laplacian were calculated, see the [[Laplacian]] article.
    
=Optimization with Constraints: Langrage Multipliers=
 
=Optimization with Constraints: Langrage Multipliers=
3,552

edits