Changes

Jump to navigation Jump to search
23 bytes added ,  21:36, February 5, 2010
m
Line 142: Line 142:  
Let <math>\vec{p}=(x_p,y_p)</math> be some point on <math>g(x,y)=c</math>.Note that <math>\nabla f (\vec{p})</math> and <math>\nabla g(\vec{p})</math> do ''not'' point along the same line, then we can always travel along <math>g(x,y)=c</math> in the direction of <math>\nabla f (\vec{p})</math> to get values for ''f'' larger than <math>f(\vec{p})</math>, or in the opposite direction of <math>\nabla f (\vec{p})</math> to get values for ''f'' smaller than <math>f(\vec{p})</math>.  Hence, <math>\vec{p}</math> cannot be a place where ''f'' achieves a local extrema on ''g'' if <math>\nabla f(\vec{p})</math> and <math>\nabla g(\vec{p})</math> do not point along the same line.  The logical conclusion, then, is that the only points <math>\vec{u}</math> on <math>g(x,y)=c</math> that could possibly be local extrema for ''f'' are those points <math>\vec{u}</math> where <math>\nabla f(\vec{u}) \propto \nabla g(\vec{u})</math>, or, stated another way, there exists some <math>\lambda</math> for which <math>\nabla f(x_0,y_0) = \lambda \nabla g(x_0,y_0)</math>.
 
Let <math>\vec{p}=(x_p,y_p)</math> be some point on <math>g(x,y)=c</math>.Note that <math>\nabla f (\vec{p})</math> and <math>\nabla g(\vec{p})</math> do ''not'' point along the same line, then we can always travel along <math>g(x,y)=c</math> in the direction of <math>\nabla f (\vec{p})</math> to get values for ''f'' larger than <math>f(\vec{p})</math>, or in the opposite direction of <math>\nabla f (\vec{p})</math> to get values for ''f'' smaller than <math>f(\vec{p})</math>.  Hence, <math>\vec{p}</math> cannot be a place where ''f'' achieves a local extrema on ''g'' if <math>\nabla f(\vec{p})</math> and <math>\nabla g(\vec{p})</math> do not point along the same line.  The logical conclusion, then, is that the only points <math>\vec{u}</math> on <math>g(x,y)=c</math> that could possibly be local extrema for ''f'' are those points <math>\vec{u}</math> where <math>\nabla f(\vec{u}) \propto \nabla g(\vec{u})</math>, or, stated another way, there exists some <math>\lambda</math> for which <math>\nabla f(x_0,y_0) = \lambda \nabla g(x_0,y_0)</math>.
    +
By solving this equation for <math>x_0,y_0,z_0</math>, we can plug our values for <math>x_0,y_0,z_0</math> into ''f'' and get the value of ''f'' at this point.  We can then check that it is an extrema, and find which kind.
   −
By solving this equation for <math>x_0,y_0,z_0</math>, we can plug our values for <math>x_0,y_0,z_0</math> into ''f'' and get the value of ''f'' at this point.  We can then check that it is an extrema, and find which kind.
+
==An Example Problem==
    
=Problems=
 
=Problems=
3,552

edits

Navigation menu