Name:___________________

VC.03 The Gradient

Literacy
 − x2 y2 
 + 
 2 8 
 
1. Calculate the gradient, ∇f [ x, y ] = gradf [ x, y ] , of f [ x, y ] = e .

2
2. Calculate the gradient, ∇f [ x, y ] = gradf [ x, y ] , of f [ x, y ] = e −4 x y .

3. Calculate the gradient, ∇f [ x, y, z ] = gradf [ x, y, z ] , of f [ x, y, z ] = xySin[ xyz ] .

 0.6
4. Go with f [ x, y ] = and look at a plot of gradf[x, y] for some selected points {x, y}
1 + 1.2 x 2 + 1.4 y 2
in the vicinity of {0, 0}. Why are these gradient vectors so attracted to the point {0, 0}?
y
1.0

0.5

x
1.0 0.5 0.5 1.0

0.5

( )
1.0

5. Go with f [ x, y ] = 0.6 1 + 1.2 x 2 + 1.4 y 2 and look at a plot of gradf[x, y] for some selected points
{x, y} in the vicinity of {0, 0}. Why are these gradient vectors so repelled by the point {0, 0}?
y

x
2 1 1 2

1

2

6. Given a certain function f[x, y], you find that the plot of gradf[x, y] with the tails at {x, y} for some
selected points {x, y} in the vicinity of {0, 0} looks like this:
y
1.0

0.5

x
1.5 1.0 0.5 0.5 1.0 1.5

0.5

1.0

Explain why the plot suggests that when you go with a point {x, y} with –1 < x < 1 and –1 < y < 1,
then you can expect that:

a) If x > 0, then increasing x results in increasing f[x, y]

b) If y > 0, then increasing y results in decreasing f[x, y]

7. You are at the point {2, 0}
a) In the direction of what vector should you step off {2, 0} in order to get the greatest initial
increase of the function f [ x, y ] = x 2 + 7 xe y ?

b) b) In the direction of what vector should you step off {2, 0} in order to get the greatest
initial decrease of the function f [ x, y ] = x 2 + 7 xe y ?

8. Go with f [ x, y, z ] = xyCos[ z ]
a) In which direction should you leave the point {1, 3, 0} to get the greatest possible initial
increase of f [x, y, z]?

b) In which direction should you leave the point {1, 3, 0} to get the greatest possible initial
decrease of f [x, y, z]?

9. The point {1, 1} is on the level curve x 2 y 3 e x + y = 7.38906 . Here is a plot of this level curve.

Pencil in a vector with its tail at {1, 1} that points

In the same direction as gradf [1, 1].
 2
 y
10. Here is a level surface f [ x, y, z ] = x +   + z 2 = 1 together with a selection of gradient vectors
2

2
of f [x, y, z] with its tails at the points at which they are calculated:

a) Explain why you are not surprised that the plotted gradient vectors
are perpendicular to the level surface.

b) Explain why these gradient vectors point out away from the surface and not in toward the interior of
the surface.

11. For the following, you must give clear, detailed answers based on your analysis of the function.
THIS IS VERY IMPORTANT!

a) Does f [ x, y ] = x 6 − 3 x 2 y + y 4 − 6 x + 5 y have a maximizer or minimizer? EXPLAIN!

( )
b) Does f [ x, y ] = x 6 + 7 y 3 + 4 e − x
2 − y2
have a maximizer or minimizer? EXPLAIN!
 x5 + 3y 2
c) Does f [ x, y ] = have a maximizer or minimizer? EXPLAIN!
1+ x4 + y6

12. You are sitting at a given point {x0 , y 0 , z 0 } , and are trying to study a function f [x, y, z]. As you
leave this point in the direction of a vector U, is it true that the function f [x, y, z] initially goes up if
U .gradf [ x0 , y 0 , z 0 ] > 0 and initially goes down if U .gradf [ x0 , y 0 , z 0 ] < 0 ?

13. When you are trying to find the highest crest or the deepest dip on the plot of a surface z = f [x, y],
why do you become interested in solving gradf [x, y] = {0, 0} for x and y?

14. Use the total differential of f [x, y] = xy to give a quick hand derivation of the “product rule”.
D [ x[t ], y[t ], t ] = x ′ [t ] y[t ] + x[t ] y ′ [t ]
 x
15a. Use the total differential of f [ x, y ] = to give a quick hand derivation of the “quotient rule”.
y
 x[t ]  x ′ [t ] x[t ] y ′ [t ] x ′ [t ] y[t ] − x[t ] y ′ [t ]
D ,t = − =
 y[t ]  y[t ] y[t ] 2 y[t ] 2

15b. Prove that gradient vectors are always perpendicular to a level curve.

∂f [ x, y ]
16. Lots of folks like to use the notations f (1,0) [ x, y ] = = D[ f [ x, y ], x] and
∂x
∂f [ x, y ] 2 2
f ( 0,1) [ x, y ] = = D[ f [ x, y ], y ] . If f [ x, y ] = e x −2 y …
∂y

∂f [ x, y ]
a) Calculate f (1,0) [ x, y ] =
∂x

∂f [ x, y ]
b) Calculate f ( 0,1) [ x, y ] =
∂y
 ∂f [ x, y, z ]
17. Calculate f ( 0,0,1) [ x, y, z ] = for f [ x, y, z ] = Cos[ x 2 y 4 z 3 ]
∂z

∂f [2.3, 8.4]
18. If f (1,0) [2.3, 8.4] = = −0.72017 for some function f[x, y], then what does this
∂x
calculation tell you about what happens to f[x, y] when you hold y at 8.4, but you make x just a teensy-
bit bigger than 2.3? Explain.

∂f [ x, y ] ∂f [ x, y ]
19. a) If f [x, y] = Sin[2x – 3y], calculate 3 +2
∂x ∂y

b) When you go with any function g(x) and you put f [x, y] = g [2x – 3y], calculate
∂f [ x, y ] ∂f [ x, y ]
3 +2
∂x ∂y
20) You are trying to minimize a function f [x, y, z] subject to the constraint that {x, y, z} must lie on a
given line in 3D. Explain why you want to become very interested in points on the line at which
∇f [ x , y , z ] =
gradf [ x, y, z ] is perpendicular to the line.

21. Use the fundamental theorem of calculus and the chain rule

D[ f [ x[t ], y[t ]], t ] = gradf [ x[t ], y[t ]].{x '[t ], y '[t ]}

for functions of one variable to give the clean formula for the integral

∫ gradf [ x[t ], y[t ]].{x '[t ], y '[t ]}dt

a

22. If you are given that a = 0, b = 2π, and

x[t ] = Cos[t ]

y[t ] = Sin[t ]

explain how you also know that

∫ gradf [ x[t ], y[t ]].{x '[t ], y '[t ]}dt = 0

a