| 
   
EXAMPLES FOR
SECTION 13.6 
  
  
    | Directional Derivative Example
       If your answer to the question posed at the end
      of the Section 13.3 #40  example related to the directional derivative, or if you have
      not yet heard about directional derivatives but your answer somehow
      related to a derivative as a rate of change relative to change along the
      line y = (4/3)x or change relative to change in the direction of a vector,
      say <3,4>, then good show.  Here is the DPGraph
Picture  again that the question related to.  Maple Picture 
        
      Here are some notes
      on the directional derivative and more examples involving the gradient
      and the directional derivative..  
     | 
   
 
  
 
  
  
    | Section 13.6 #32 Maximizing the Directional Derivative 
      DPGraph
      Picture of the Surface     DPGraph
      Picture with the Point 
      
      DPGraph
      Picture of the surface, the point, and a plane containing the point
      and a line in the xy-coordinate plane with direction vector u where
      the x- and y-components of u equal the x- and y-components of the
      direction vector for the directional derivative.  See the analysis
      below for more details.
        
  | 
   
 
  
 
  
	
		| Maximizing the Directional 
		Derivative Find the 
		maximum value of the directional derivative of the given function at the 
		indicated point. 
		 
  | 
		
		 
		Animation  
		QT Version  | 
	 
 
   
  
Powerpoint
Presentation on the Relationship Between the Gradient and Level Curves  
  
 
  
  
    | 
 Heat
Seeking Particle Examples 
T(x,y) = 100 - x2 - 2y2       
Starting Point (2,4) 
Let the path be described by  r(t) = <
x(t) , y(t) > 
Then  r'(t) = < dx/dt , dy/dt > 
gives the direction of motion. 
      Since this is a heat seeking particle it must move at
all times in the direction of maximum increase in temperature.  This must
be the direction of the gradient of the temperature function T(x,y). 
      grad T(x,y) = < -2x , -4y >         
		Powerpoint Enhancement  | 
    
	 
      Click here to see a
      Maple worksheet on this example. 
	Click on the picture to see a larger picture of the 
	analytical solution constrained so that the path looks orthogonal to the 
	equal temperature lines.  | 
   
 
In general then it must be the case that grad T =
k r'(t) where k is some constant.  Let us assume a parameterization
of the path such that k = 1 so that grad T = r'(t).  Note
that we are not claiming that r'(t) gives the velocity of the heat
seeking particle, only that it gives the direction. 
From grad T = r'(t) we equate vector
components to get dx/dt = -2x and dy/dt = -4y.  Let 0 be the value of t at
the starting point so that x(0) = 2 and y(0) = 4.  Thus we have two
differential equations to solve. 
  
    | 
       dx/dt =
      -2x       x(0) = 2 
      (1/x)dx =
      -2dt 
      ln | x |
      = -2t + c 
      | x | = e-2t
      + c = e-2t ec = c1e-2t 
      x = c2e-2t 
      x = 2e-2t  
      since x(0) = 2 yields c2 = 2 
        
      In a
      similar fashion we would find that from 
      dy/dt =
      -4y and y(0) = 4 we get 
        
      y = 4e-4t 
        
      Thus the
      path is given by r(t) = < 2e-2t , 4e-4t
      >. 
      Since y =
      (2e-2t)2 we see that the path is also given by y = x2. 
        
      The
      picture on the right shows the path.  Animation  | 
    
      | 
   
 
A similar problem involves finding the path followed by
a heat seeking object in space where the temperature distribution is given by 
 T(x,y,z) = 100 - 3x - y - z2 and the starting point is (2,2,5).
 Let the path be described by  r(t) = <
x(t) , y(t) , z(t) > 
Then  r'(t) = < dx/dt , dy/dt , dz/dt > 
gives the direction of motion. 
Since this is a heat seeking particle it must move at
all times in the direction of maximum increase in temperature.  This must
be the direction of the gradient of the temperature function T(x,y,z). 
grad T(x,y,z) = < -3 , -1 , -2z > 
  
    | 
 In general then it must be the case that grad T =
k r'(t) where k is some constant.  Let us assume a parameterization
of the path such that k = 1 so that grad T = r'(t).  Note
that we are not claiming that r'(t) gives the velocity of the heat
seeking particle, only that it gives the direction. 
From grad T = r'(t) we equate vector
components to get  
dx/dt = -3, dy/dt = -1, and
dz/dt = -2z.   
 Let 0 be the value of t at
the starting point so that  
 x(0) = 2, y(0) = 2, and z(0) = 5.   
We have
three
differential equations to solve and their solutions  
 will yield x = -3t + 2, y =
-t + 2, and z = 5e-2t. 
Thus r(t) = < -3t + 2 , -t + 2 , 5e-2t
> 
      Click here to see an
      animated
picture of a portion of the path.  | 
      | 
   
 
  
 
Heat Seeking Particle Problem To Practice On 
Find the path followed by a heat seeking particle on a
plate whose temperature distribution is given by T(x,y) starting at the given
point (two cases given). 
  
     
      Click
      here to see a Maple worksheet on the problems above.  | 
      | 
   
 
  
 
The Billy Johnson Heat Seeking Particle Problem 
The idea for this problem was suggested by Billy 
Johnson.  The temperature distribution function given below has multiple 
relative hot and relative cold spots.  Finding the path followed by a heat 
seeking particle (or object) starting at some given point required using the 
gradient repeatedly to approximate the direction of motion.  The linked 
Maple Worksheet shows many possibilities and how Maple can approximate the 
paths. 
                           
Maple Worksheet 
 
Section 13.6 #76  This is a technology 
problem--see this Maple Worksheet 
 
Click here to see a Maple
worksheet depicting a skier taking the path of steepest descent down a mountain. 
Here is my Maple Twin Peaks worksheet. 
 
  
return
  |