You are on page 1of 37

Strauss PDEs 2e: Section 1.

1 - Exercise 1 Page 1 of 8

Exercise 1
Verify the linearity and nonlinearity of the eight examples of PDEs given in the text, by checking
whether or not equations (3) are valid.

Solution

The eight examples in the text are given below.

1. ux + uy = 0 (transport)
2. ux + yuy = 0 (transport)
3. ux + uuy = 0 (shock wave)
4. uxx + uyy = 0 (Laplace’s equation)
5. utt − uxx + u3 = 0 (wave with interaction)
6. ut + uux + uxxx = 0 (dispersive wave)
7. utt + uxxxx = 0 √ (vibrating bar)
8. ut − iuxx = 0 (i = −1) (quantum mechanics)
The equations in (3) are L (u + v) = L u + L v and L (cu) = cL u, which state the conditions for
linearity. They will be checked in each case.

Example One:

ux + uy = 0 (transport)
∂ ∂
u+ u=0 (change notation)
∂x ∂y
 
∂ ∂
+ u=0 (factor the common term)
∂x ∂y
Lu = 0

Therefore, for this example, the operator we are working with is


∂ ∂
L = +
∂x ∂y
Now the equations in (3) may be verified. Suppose u and v are solutions to the PDE. The first
equation in (3) yields
 
∂ ∂
L (u + v) = + (u + v)
∂x ∂y
∂ ∂
= (u + v) + (u + v)
∂x ∂y
∂ ∂ ∂ ∂
= u+ v+ u+ v
∂x ∂x ∂y ∂y
∂ ∂ ∂ ∂
= u+ u+ v+ v
∂x ∂y ∂x ∂y
   
∂ ∂ ∂ ∂
= + u+ + v
∂x ∂y ∂x ∂y
= Lu+Lv

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 1 Page 2 of 8

Thus, the first equation in (3) is satisfied. Now for the second equation in (3).

 
∂ ∂
L (cu) = + (cu)
∂x ∂y
∂ ∂
= (cu) + (cu)
∂x ∂y
∂ ∂
=c u+c u
∂x ∂y
 
∂ ∂
=c u+ u
∂x ∂y
 
∂ ∂
=c + u
∂x ∂y
= cL u

The second equation in (3) is also satisfied, and hence the PDE in this example is linear.

Example Two:

ux + yuy = 0 (transport)
∂ ∂
u+y u=0 (change notation)
∂x ∂y
 
∂ ∂
+y u=0 (factor the common term)
∂x ∂y
Lu = 0

Therefore, for this example, the operator we are working with is


∂ ∂
L = +y
∂x ∂y

Now the equations in (3) may be verified. Suppose u and v are solutions to the PDE. The first
equation in (3) yields
 
∂ ∂
L (u + v) = +y (u + v)
∂x ∂y
∂ ∂
= (u + v) + y (u + v)
∂x ∂y
∂ ∂ ∂ ∂
= u+ v+y u+y v
∂x ∂x ∂y ∂y
∂ ∂ ∂ ∂
= u+y u+ v+y v
∂x ∂y ∂x ∂y
   
∂ ∂ ∂ ∂
= +y u+ +y v
∂x ∂y ∂x ∂y
= Lu+Lv

Thus, the first equation in (3) is satisfied. Now for the second equation in (3).

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 1 Page 3 of 8

 
∂ ∂
L (cu) = +y (cu)
∂x ∂y
∂ ∂
= (cu) + y (cu)
∂x ∂y
∂ ∂
= c u + cy u
∂x ∂y
 
∂ ∂
=c u+y u
∂x ∂y
 
∂ ∂
=c +y u
∂x ∂y
= cL u

The second equation in (3) is also satisfied, and hence the PDE in this example is linear.

Example Three:

ux + uuy = 0 (shock wave)


∂ ∂
u + u u = 0 (change notation)
∂x ∂y
 
∂ ∂
+u u = 0 (factor the common term)
∂x ∂y
Lu = 0

Therefore, for this example, the operator we are working with is


∂ ∂
L = +u
∂x ∂y
Now the equations in (3) may be verified. Suppose u and v are solutions to the PDE. The first
equation in (3) yields
 
∂ ∂
L (u + v) = + (u + v) (u + v)
∂x ∂y
∂ ∂
= (u + v) + (u + v) (u + v)
∂x ∂y
 
∂ ∂ ∂ ∂
= u+ v + (u + v) u+ v
∂x ∂x ∂y ∂y
∂ ∂ ∂ ∂ ∂ ∂
= u+ v+u u+u v+v u+v v
∂x ∂x ∂y ∂y ∂y ∂y
∂ ∂ ∂ ∂ ∂ ∂
= u+u u+ v+v v+u v+v u
∂x ∂y ∂x ∂y ∂y ∂y
   
∂ ∂ ∂ ∂ ∂ ∂
= +u u+ +v v+u v+v u
∂x ∂y ∂x ∂y ∂y ∂y
= L u + L v + uvy + vuy

Thus, the first equation in (3) is not satisfied. The PDE in this example is nonlinear.

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 1 Page 4 of 8

Example Four:

uxx + uyy = 0 (Laplace’s equation)


∂2 ∂2
u + 2 u = 0 (change notation)
∂x2 ∂y
∂2 ∂2
 
+ u = 0 (factor the common term)
∂x2 ∂y 2
Lu = 0

Therefore, for this example, the operator we are working with is

∂2 ∂2
L = +
∂x2 ∂y 2

Now the equations in (3) may be verified. Suppose u and v are solutions to the PDE. The first
equation in (3) yields
 2
∂2


L (u + v) = + (u + v)
∂x2 ∂y 2
∂2 ∂2
= 2
(u + v) + 2 (u + v)
∂x ∂y
∂ 2 ∂ 2 ∂2 ∂2
= u + v + u + v
∂x2 ∂x2 ∂y 2 ∂y 2
∂2 ∂2 ∂2 ∂2
= 2
u + 2u + 2v + 2v
∂x ∂y ∂x ∂y
 2 2
 2
∂2
 
∂ ∂ ∂
= + u+ + v
∂x2 ∂y 2 ∂x2 ∂y 2
= Lu+Lv

Thus, the first equation in (3) is satisfied. Now for the second equation in (3).

∂2 ∂2
 
L (cu) = + (cu)
∂x2 ∂y 2
∂2 ∂2
= (cu) + (cu)
∂x2 ∂y 2
∂2 ∂2
= c 2u + c 2u
∂x ∂y
 2
∂2


=c u + 2u
∂x2 ∂y
 2
∂2


=c + u
∂x2 ∂y 2
= cL u

The second equation in (3) is also satisfied, and hence the PDE in this example is linear.

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 1 Page 5 of 8

Example Five:

utt − uxx + u3 = 0 (wave with interaction)


∂2 ∂2
2
u − 2 u + u3 = 0 (change notation)
∂t
 2 ∂x
∂2

∂ 2
− + u u = 0 (factor the common term)
∂t2 ∂x2
Lu = 0

Therefore, for this example, the operator we are working with is

∂2 ∂2
L = − + u2
∂t2 ∂x2
Now the equations in (3) may be verified. Suppose u and v are solutions to the PDE. The first
equation in (3) yields
 2
∂2


L (u + v) = − 2
+ (u + v) (u + v)
∂t2 ∂x2
∂2 ∂2
= 2 (u + v) − 2 (u + v) + (u + v)3
∂t ∂x
∂2 ∂2 ∂2 ∂2
= 2 u + 2 v − 2 u − 2 v + u3 + u2 v + uv 2 + v 3
∂t ∂t ∂x ∂x
∂2 ∂2 ∂ 2 ∂2
= 2 u − 2 u + u3 + 2 v − 2 v + v 3 + u2 v + uv 2
∂t ∂x  ∂t 2 ∂x 2
∂2 ∂2
 
2 ∂ ∂
= − +u u+ − + v v + u2 v + uv 2
2
∂t2 ∂x2 ∂t2 ∂x2
= L u + L v + u2 v + uv 2

Thus, the first equation in (3) is not satisfied. The PDE in this example is nonlinear.

Example Six:

ut + uux + uxxx = 0 (dispersive wave)


∂ ∂ ∂3
u + u u + 3u = 0 (change notation)
∂t ∂x ∂x
∂ ∂ ∂3
+u + u=0 (factor the common term)
∂t ∂x ∂x3
Lu = 0

Therefore, for this example, the operator we are working with is

∂ ∂ ∂3
L = +u + 3
∂t ∂x ∂x
Now the equations in (3) may be verified. Suppose u and v are solutions to the PDE. The first

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 1 Page 6 of 8

equation in (3) yields

∂3
 
∂ ∂
L (u + v) = + (u + v) + (u + v)
∂t ∂x ∂x3
∂ ∂ ∂3
= (u + v) + (u + v) (u + v) + 3 (u + v)
∂t ∂x ∂x
∂3 ∂3
 
∂ ∂ ∂ ∂
= u + v + (u + v) u+ v + 3u + 3v
∂t ∂t ∂x ∂x ∂x ∂x
∂ ∂ ∂ ∂ ∂ ∂ ∂3 ∂3
= u + v + u u + u v + v u + v v + 3u + 3v
∂t ∂t ∂x ∂x ∂x ∂x ∂x ∂x
∂ ∂ ∂3 ∂ ∂ ∂3 ∂ ∂
= u + u u + 3u + v + v v + 3v + u v + v u
∂t ∂x ∂x  ∂t ∂x ∂x  ∂x ∂x
∂3 ∂3
 
∂ ∂ ∂ ∂ ∂ ∂
= +u + 3 u+ +v + 3 v+u v+v u
∂t ∂x ∂x ∂t ∂x ∂x ∂x ∂x
= L u + L v + uvx + vux

Thus, the first equation in (3) is not satisfied. The PDE in this example is nonlinear.

Example Seven:

utt + uxxxx = 0 (vibrating bar)


∂2 ∂4
2
u + 4u = 0 (change notation)
 ∂t ∂x
∂2 ∂4
+ u=0 (factor the common term)
∂t2 ∂x4
Lu = 0

Therefore, for this example, the operator we are working with is

∂2 ∂4
L = 2
+ 4
∂t ∂x
Now the equations in (3) may be verified. Suppose u and v are solutions to the PDE. The first
equation in (3) yields
 2
∂4


L (u + v) = + (u + v)
∂t2 ∂x4
∂2 ∂4
= 2 (u + v) + 4 (u + v)
∂t ∂x
∂2 ∂2 ∂4 ∂4
= 2u + 2v + 4u + 4v
∂t ∂t ∂x ∂x
∂2 ∂4 ∂2 ∂4
= 2u + 4u + 2v + 4v
∂t ∂x  ∂t ∂x
∂2 ∂4 ∂2 ∂4
 
= + u+ + v
∂t2 ∂x4 ∂t2 ∂x4
= Lu+Lv

Thus, the first equation in (3) is satisfied. Now for the second equation in (3).

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 1 Page 7 of 8

∂2 ∂4
 
L (cu) = + (cu)
∂t2 ∂x4
∂2 ∂4
= 2 (cu) + 4 (cu)
∂t ∂x
∂2 ∂4
= c 2u + c 4u
∂t
 2 ∂x
∂4


=c u + 4u
∂t2 ∂x
 2
∂4


=c + u
∂t2 ∂x4
= cL u

The second equation in (3) is also satisfied, and hence the PDE in this example is linear.

Example Eight:

ut − iuxx = 0 (i = −1) (transport)
∂ ∂2
u − i 2u = 0 (change notation)
 ∂t ∂x
∂ ∂2
−i 2 u=0 (factor the common term)
∂t ∂x
Lu = 0

Therefore, for this example, the operator we are working with is

∂ ∂2
L = −i 2
∂t ∂x
Now the equations in (3) may be verified. Suppose u and v are solutions to the PDE. The first
equation in (3) yields

∂2
 

L (u + v) = − i 2 (u + v)
∂t ∂x
∂ ∂2
= (u + v) − i 2 (u + v)
∂t ∂x
∂ ∂ ∂2 ∂2
= u + v − i 2u − i 2v
∂t ∂t ∂x ∂x
∂ ∂2 ∂ ∂2
= u − i 2u + v − i 2v
∂t ∂x  ∂t ∂x
∂2 ∂2
 
∂ ∂
= −i 2 u+ −i 2 v
∂t ∂x ∂t ∂x
= Lu+Lv

Thus, the first equation in (3) is satisfied. Now for the second equation in (3).

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 1 Page 8 of 8

∂2
 

L (cu) = − i 2 (cu)
∂t ∂x
∂ ∂2
= (cu) − i 2 (cu)
∂t ∂x
∂ ∂2
= c u − ic 2 u
∂t ∂x
∂2
 

=c u − i 2u
∂t ∂x
∂2
 

=c −i 2 u
∂t ∂x
= cL u

The second equation in (3) is also satisfied, and hence the PDE in this example is linear.

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 2 Page 1 of 5

Exercise 2
Which of the following operators are linear?

(a) L u = ux + xuy

(b) L u = ux + uuy

(c) L u = ux + u2y

(d) L u = ux + uy + 1

(e) L u = 1 + x2 (cos y)ux + uyxy − [arctan (x/y)]u

Solution

To determine if an operator is linear, one must check whether the conditions for linearity hold:

L (u + v) = L u + L v and L (cu) = cL u

Part (a)

L u = ux + xuy
∂ ∂
Lu = u+x u
∂x ∂y
 
∂ ∂
= +x u
∂x ∂y

Therefore,
∂ ∂
L = +x
∂x ∂y
 
∂ ∂
L (u + v) = +x (u + v)
∂x ∂y
∂ ∂
= (u + v) + x (u + v)
∂x ∂y
∂ ∂ ∂ ∂
= u+ v+x u+x v
∂x ∂x ∂y ∂y
∂ ∂ ∂ ∂
= u+x u+ v+x v
∂x ∂y ∂x ∂y
   
∂ ∂ ∂ ∂
= +x u+ +x v
∂x ∂y ∂x ∂y
= Lu+Lv

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 2 Page 2 of 5

The first condition holds. Now the second one must be checked.
 
∂ ∂
L (cu) = +x (cu)
∂x ∂y
∂ ∂
= (cu) + x (cu)
∂x ∂y
∂ ∂
= c u + cx u
∂x ∂y
 
∂ ∂
=c u+x u
∂x ∂y
 
∂ ∂
=c +x u
∂x ∂y
= cL u

The second condition holds as well, so the operator is linear.

Part (b)

L u = ux + uuy
∂ ∂
Lu = u+u u
∂x ∂y
 
∂ ∂
= +u u
∂x ∂y

Therefore,
∂ ∂
L = +u
∂x ∂y
 
∂ ∂
L (u + v) = + (u + v) (u + v)
∂x ∂y
∂ ∂
= (u + v) + (u + v) (u + v)
∂x ∂y
 
∂ ∂ ∂ ∂
= u+ v + (u + v) u+ v
∂x ∂x ∂y ∂y
∂ ∂ ∂ ∂ ∂ ∂
= u+ v+u u+u v+v u+v v
∂x ∂x ∂y ∂y ∂y ∂y
∂ ∂ ∂ ∂ ∂ ∂
= u+u u+ v+v v+u v+v u
∂x ∂y ∂x ∂y ∂y ∂y
   
∂ ∂ ∂ ∂ ∂ ∂
= +u u+ +v v+u v+v u
∂x ∂y ∂x ∂y ∂y ∂y
= L u + L v + uvy + vuy

The first condition does not hold, so the PDE is nonlinear.

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 2 Page 3 of 5

Part (c)

L u = ux + u2y
 2
∂ ∂
Lu = u+ u
∂x ∂y
   
∂ ∂ ∂
= + u u
∂x ∂y ∂y

Therefore,  
∂ ∂ ∂
L = + u
∂x ∂y ∂y
   
∂ ∂ ∂
L (u + v) = + (u + v) (u + v)
∂x ∂y ∂y
 
∂ ∂ ∂
= (u + v) + (u + v) (u + v)
∂x ∂y ∂y
  
∂ ∂ ∂ ∂ ∂ ∂
= u+ v+ u+ v u+ v
∂x ∂x ∂y ∂y ∂y ∂y
 2        2
∂ ∂ ∂ ∂ ∂ ∂ ∂ ∂
= u+ v+ u + u v + v u + v
∂x ∂x ∂y ∂y ∂y ∂y ∂y ∂y
 2  2   
∂ ∂ ∂ ∂ ∂ ∂
= u+ u + v+ v +2 u v
∂x ∂y ∂x ∂y ∂y ∂y
          
∂ ∂ ∂ ∂ ∂ ∂ ∂ ∂
= + u u+ + v v+2 u v
∂x ∂y ∂y ∂x ∂y ∂y ∂y ∂y
= L u + L v + 2uy vy

The first condition does not hold, so the PDE is nonlinear.

Part (d)

L u = ux + uy + 1
∂ ∂
Lu = u+ u+1
∂x ∂y
 
∂ ∂ 1
= + + u
∂x ∂y u

Therefore,
∂ ∂ 1
L = + +
∂x ∂y u

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 2 Page 4 of 5

 
∂ ∂ 1
L (u + v) = + + (u + v)
∂x ∂y u + v
∂ ∂
= (u + v) + (u + v) + 1
∂x ∂y
∂ ∂ ∂ ∂
= u+ v+ u+ v+1
∂x ∂x ∂y ∂y
∂ ∂ ∂ ∂
= u+ u+1+ v+ v+1−1
∂x ∂y ∂x ∂y
   
∂ ∂ 1 ∂ ∂ 1
= + + u+ + + v−1
∂x ∂y u ∂x ∂y v
= Lu+Lv −1

The first condition does not hold, so the PDE is nonlinear.

Part (e)

p
Lu = 1 + x2 (cos y)ux + uyxy − [arctan (x/y)]u
3
 
p ∂ ∂ x
L u = 1 + x2 (cos y) u + u − arctan u
∂x ∂y∂x∂y y
∂3
p 
2
∂ x
= 1 + x (cos y) + − arctan u
∂x ∂y∂x∂y y

Therefore,
p ∂ ∂3 x
L = 1 + x2 (cos y) + − arctan
∂x ∂y∂x∂y y

∂3
p 
∂ x
L (u + v) = 2
1 + x (cos y) + − arctan (u + v)
∂x ∂y∂x∂y y
∂3
 
p
2
∂ x
= 1 + x (cos y) (u + v) + (u + v) − arctan (u + v)
∂x ∂y∂x∂y y
∂3 ∂3
     
p
2
∂ ∂ x
= 1 + x (cos y) u+ v + u+ v − arctan (u + v)
∂x ∂x ∂y∂x∂y ∂y∂x∂y y
∂3
p 
2
∂ x
= 1 + x (cos y) + − arctan u
∂x ∂y∂x∂y y
3
p 
∂ ∂ x
+ 1 + x2 (cos y) + − arctan v
∂x ∂y∂x∂y y
= Lu+Lv

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 2 Page 5 of 5

The first condition holds. Now the second one must be checked.
∂3
p 
∂ x
L (cu) = 1 + x2 (cos y) + − arctan (cu)
∂x ∂y∂x∂y y
∂3
 
p
2
∂ x
= 1 + x (cos y) (cu) + (cu) − arctan (cu)
∂x ∂y∂x∂y y
∂3
 
p
2
∂ x
= c 1 + x (cos y) u + c u − c arctan u
∂x ∂y∂x∂y y
∂3
p   
∂ x
=c 1 + x2 (cos y) u + u − arctan u
∂x ∂y∂x∂y y
∂3
p 
2
∂ x
=c 1 + x (cos y) + − arctan u
∂x ∂y∂x∂y y
= cL u

The second condition holds as well, so the operator is linear.

To summarize, (a) and (e) are linear, whereas (b), (c), and (d) are nonlinear.

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 3 Page 1 of 7

Exercise 3
For each of the following equations, state the order and whether it is nonlinear, linear
inhomogeneous, or linear homogeneous; provide reasons.
(a) ut − uxx + 1 = 0
(b) ut − uxx + xu = 0
(c) ut − uxxt + uux = 0
(d) utt − uxx + x2 = 0
(e) iut − uxx + u/x = 0
(f) ux (1 + u2x )−1/2 + uy (1 + u2y )−1/2 = 0
(g) ux + ey uy = 0

(h) ut + uxxxx + 1 + u = 0

Solution

The order of an equation is the highest derivative that appears. To determine if an operator is
linear, one must check whether the conditions for linearity hold:
L (u + v) = L u + L v and L (cu) = cL u
Assuming that L is a linear operator, the equation L u = 0 is a homogeneous linear equation,
and the equation L u = g (g 6= 0) is an inhomogeneous linear equation.

Part (a)

ut − uxx + 1 = 0
This PDE is of the second order because the second derivative is the highest derivative.

The equation can be written as


∂ ∂2
L u = −1, where L = − 2
∂t ∂x
Checking the first condition,
∂2
 

L (u + v) = − 2 (u + v)
∂t ∂x
∂ ∂2
= (u + v) − 2 (u + v)
∂t ∂x
∂ ∂ ∂2 ∂2
= u + v − 2u − 2v
∂t ∂t ∂x ∂x
∂ ∂ 2 ∂ ∂2
= u − 2u + v − 2v
∂t ∂x  ∂t ∂x
∂2 ∂2
 
∂ ∂
= − u+ − v
∂t ∂x2 ∂t ∂x2
= Lu+Lv

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 3 Page 2 of 7

The first condition for linearity holds. Now the second one must be checked.
∂2
 

L (cu) = − (cu)
∂t ∂x2
∂ ∂2
= (cu) − 2 (cu)
∂t ∂x
∂ ∂2
= c u − c 2u
∂t ∂x 
∂2


=c − 2 u
∂t ∂x
= cL u
The second condition for linearity is satisfied as well, so the PDE is a linear inhomogeneous one.

Part (b)

ut − uxx + xu = 0
This PDE is of the second order because the second derivative is the highest derivative.

The equation can be written as


∂ ∂2
L u = 0, where L = − 2 +x
∂t ∂x
Checking the first condition,
∂2
 

L (u + v) = − + x (u + v)
∂t ∂x2
∂ ∂2
= (u + v) − 2 (u + v) + x(u + v)
∂t ∂x
∂ ∂ ∂2 ∂2
= u + v − 2 u − 2 v + xu + xv
∂t ∂t ∂x ∂x
∂ ∂2 ∂ ∂2
= u − 2 u + xu + v − 2 v + xv
∂t ∂x ∂t ∂x
∂2 ∂2
   
∂ ∂
= − +x u+ − +x v
∂t ∂x2 ∂t ∂x2
= Lu+Lv
The first condition for linearity holds. Now the second one must be checked.
∂2
 

L (cu) = − + x (cu)
∂t ∂x2
∂ ∂2
= (cu) − 2 (cu) + x(cu)
∂t ∂x
∂ ∂2
= c u − c 2 u + cxu
∂t ∂x
∂2
 

=c − +x u
∂t ∂x2
= cL u

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 3 Page 3 of 7

The second condition for linearity is satisfied as well, so the PDE is a linear homogeneous one.

Part (c)

ut − uxxt + uux = 0
This PDE is of the third order because the third derivative is the highest derivative.

The equation can be written as

∂ ∂3 ∂
L u = 0, where L = − 2 +u
∂t ∂x ∂t ∂x
Checking the first condition,

∂3
 
∂ ∂
L (u + v) = − + (u + v) (u + v)
∂t ∂x2 ∂t ∂x
∂ ∂3 ∂
= (u + v) − 2 (u + v) + (u + v) (u + v)
∂t ∂x ∂t ∂x 
∂3 ∂3

∂ ∂ ∂ ∂
= u + v − 2 u − 2 v + (u + v) u+ v
∂t ∂t ∂x ∂t ∂x ∂t ∂x ∂x
∂ ∂ ∂3 ∂3 ∂ ∂ ∂ ∂
= u+ v− 2 u− 2 v+u u+u v+v u+v v
∂t ∂t ∂x ∂t ∂x ∂t ∂x ∂x ∂x ∂x
∂ ∂3 ∂ ∂ ∂3 ∂ ∂ ∂
= u− 2 u+u u+ v− 2 v+v v+u v+v u
∂t ∂x ∂t ∂x ∂t ∂x ∂t ∂x  ∂x ∂x
∂3 ∂3
 
∂ ∂ ∂ ∂ ∂ ∂
= − 2 +u u+ − 2 +v v+u v+v u
∂t ∂x ∂t ∂x ∂t ∂x ∂t ∂x ∂x ∂x
= L u + L v + uvx + vux

The first condition for linearity does not hold, so the equation is nonlinear.

Part (d)

utt − uxx + x2 = 0
This PDE is of the second order because the second derivative is the highest derivative.

The equation can be written as

∂2 ∂2
L u = −x2 , where L = −
∂t2 ∂x2

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 3 Page 4 of 7

Checking the first condition,

∂2 ∂2
 
L (u + v) = − (u + v)
∂t2 ∂x2
∂2 ∂2
= (u + v) − (u + v)
∂t2 ∂x2
∂2 ∂2 ∂2 ∂2
= 2u + 2v − 2u − 2v
∂t ∂t ∂x ∂x
∂ 2 ∂ 2 ∂ 2 ∂2
= 2u − 2u + 2v − 2v
∂t ∂x  ∂t ∂x
2 2 2 ∂2
 
∂ ∂ ∂
= − u+ − v
∂t2 ∂x2 ∂t2 ∂x2
= Lu+Lv

The first condition for linearity holds. Now the second one must be checked.
 2
∂2


L (cu) = − (cu)
∂t2 ∂x2
∂2 ∂2
= 2 (cu) − 2 (cu)
∂t ∂x
∂2 ∂2
= c 2u − c 2u
∂t
 2 ∂x 
∂ ∂2
=c − u
∂t2 ∂x2
= cL u

The second condition for linearity is satisfied as well, so the PDE is a linear inhomogeneous one.

Part (e)

iut − uxx + u/x = 0


This PDE is of the second order because the second derivative is the highest derivative.

The equation can be written as

∂ ∂2 1
L u = 0, where L = i − 2+
∂t ∂x x

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 3 Page 5 of 7

Checking the first condition,


∂2
 
∂ 1
L (u + v) = i − 2 + (u + v)
∂t ∂x x
∂ ∂2 1
= i (u + v) − 2 (u + v) + (u + v)
∂t ∂x x
∂ ∂ ∂2 ∂2 1 1
= i u + i v − 2u − 2v + u + v
∂t ∂t ∂x ∂x x x
∂ ∂2 1 ∂ ∂2 1
= i u − 2u + u + i v − 2v + v
∂t ∂x x ∂t ∂x x
∂2 ∂2

∂ 1 ∂ 1
= i − 2+ u+ i − 2 + v
∂t ∂x x ∂t ∂x x
= Lu+Lv
The first condition for linearity holds. Now the second one must be checked.
∂2
 
∂ 1
L (cu) = i − 2 + (cu)
∂t ∂x x
∂ ∂2 1
= i (cu) − 2 (cu) + (cu)
∂t ∂x x
∂ ∂2 1
= ic u − c 2 u + c u
∂t ∂x x
∂ ∂2 1
=c i − 2 + u
∂t ∂x x
= cL u
The second condition for linearity is satisfied as well, so the PDE is a linear homogeneous one.

Part (f )

ux (1 + u2x )−1/2 + uy (1 + u2y )−1/2 = 0


This PDE is of the first order because the first derivative is the highest derivative.

The equation can be written as


1 ∂ 1 ∂
q u+ r u=0

 2 ∂x   2 ∂y
1+ ∂x u

1 + ∂y u
 
 1 ∂ 1 ∂ 
q +r u = 0
 ∂
 2 ∂x   2 ∂y 
1 + ∂x u ∂
1 + ∂y u

Lu = 0
The operator for this PDE is
1 ∂ 1 ∂
L =q +r

 2 ∂x   2 ∂y
1+ ∂x u

1 + ∂y u

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 3 Page 6 of 7

Checking the first condition,


 
1 ∂ 1 ∂ 
L (u + v) = 

 q ∂ 2 ∂x + r h i2 ∂y  (u + v)

1 + ∂x (u + v) ∂
1 + ∂y (u + v)
1 ∂ 1 ∂
=q (u + v) + r 2 ∂y (u + v)
∂ ∂
 2 ∂x 
1+ ∂x u + ∂x v

1 + ∂y u+ ∂
∂y v

The first condition cannot be satisfied because the square roots cannot be simplified (one for u
and one for v). Hence, the PDE is nonlinear.

Part (g)

ux + ey uy = 0
This PDE is of the first order because the first derivative is the highest derivative.

The equation can be written as


∂ ∂
L u = 0, where L = + ey
∂x ∂y
Checking the first condition,
 
∂ y ∂
L (u + v) = +e (u + v)
∂x ∂y
∂ ∂
= (u + v) + ey (u + v)
∂x ∂y
∂ ∂ ∂ ∂
= u+ v + ey u + ey v
∂x ∂x ∂y ∂y
∂ ∂ ∂ ∂
= u + ey u + v + ey v
∂x ∂y ∂x ∂y
   
∂ y ∂ ∂ y ∂
= +e u+ +e v
∂x ∂y ∂x ∂y
= Lu+Lv

The first condition for linearity holds. Now the second one must be checked.
 
∂ y ∂
L (cu) = +e (cu)
∂x ∂y
∂ ∂
= (cu) + ey (cu)
∂x ∂y
∂ ∂
= c u + cey u
∂x ∂y
 
∂ y ∂
=c +e u
∂x ∂y
= cL u

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 3 Page 7 of 7

The second condition for linearity is satisfied as well, so the PDE is a linear homogeneous one.

Part (h)

ut + uxxxx + 1+u=0
This PDE is of the fourth order because the fourth derivative is the highest derivative.

The equation can be written as



∂ ∂4 1+u
L u = 0, where L = + 4+
∂t ∂x u
Checking the first condition,
p !
∂ ∂4 1 + (u + v)
L (u + v) = + 4+ (u + v)
∂t ∂x u+v

∂ ∂4 1+u+v
= (u + v) + 4 (u + v) + (u + v)
∂t ∂x u+v
∂ ∂ ∂4 ∂4 √
= u + v + 4u + 4v + 1 + u + v
∂t ∂t ∂x ∂x
The first condition cannot be satisfied because the square root cannot be simplified (one for u and
one for v). Hence, the PDE is nonlinear.

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 4 Page 1 of 1

Exercise 4
Show that the difference of two solutions of an inhomogeneous linear equation L u = g with the
same g is a solution of the homogeneous equation L u = 0.

Solution

Suppose that v and w are solutions to this inhomogeneous linear equation.

Lv = g and L w = g

Since the equation is linear, the operator L satisfies the two conditions for linearity, namely
L (u + v) = L u + L v and L (cu) = cL u. Subtracting the second equation from the first, we get

Lv −Lw = g −g
L v + L (−w) = 0
L [v + (−w)] = 0
L (v − w) = 0

Therefore, the difference of two solutions of an inhomogeneous linear equation is a solution of the
homogeneous equation.

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 5 Page 1 of 6

Exercise 5
Which of the following collections of 3-vectors [a, b, c] are vector spaces? Provide reasons.
(a) The vectors with b = 0.

(b) The vectors with b = 1.

(c) The vectors with ab = 0.

(d) All the linear combinations of the two vectors [1, 1, 0] and [2, 0, 1].

(e) All the vectors such that c − a = 2b.

Solution

For a collection of 3-vectors V to be a vector space over R, the vector addition and scalar
multiplication operations must satisfy the following ten properties:
(A1) x + y ∈ V for all x, y ∈ V.

(A2) (x + y) + z = x + (y + z) for every x, y, z ∈ V.

(A3) x + y = y + x for every x, y ∈ V.

(A4) There is an element 0 ∈ V such that x + 0 = x for every x ∈ V.

(A5) For each x ∈ V, there is an element (−x) ∈ V such that x + (−x) = 0.

(M1) αx ∈ V for all α ∈ R and x ∈ V.

(M2) (αβ)x = α(βx) for all α, β ∈ R and every x ∈ V.

(M3) α(x + y) = αx + αy for every α ∈ R and all x, y ∈ V.

(M4) (α + β)x = αx + βx for all α, β ∈ R and every x ∈ V.

(M5) 1x = x for every x ∈ V.


Part (a)

V = {[a, 0, c] | a, c ∈ R}
Choose x = [a1 , 0, c1 ] ∈ V and y = [a2 , 0, c2 ] ∈ V and z = [a3 , 0, c3 ] ∈ V.

x + y = [a1 , 0, c1 ] + [a2 , 0, c2 ] = [a1 + a2 , 0, c1 + c2 ]

Because a1 + a2 ∈ R and c1 + c2 ∈ R, x + y ∈ V. Hence, property A1 is satisfied.

(x + y) + z = ([a1 , 0, c1 ] + [a2 , 0, c2 ]) + [a3 , 0, c3 ]


= [a1 + a2 , 0, c1 + c2 ] + [a3 , 0, c3 ]
= [a1 + a2 + a3 , 0, c1 + c2 + c3 ]
= [a1 , 0, c1 ] + [a2 + a3 , 0, c2 + c3 ]
= [a1 , 0, c1 ] + ([a2 , 0, c2 ] + [a3 , 0, c3 ])
= x + (y + z)

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 5 Page 2 of 6

Property A2 is satisfied.

x + y = [a1 , 0, c1 ] + [a2 , 0, c2 ]
= [a1 + a2 , 0, c1 + c2 ]
= [a2 + a1 , 0, c2 + c1 ]
= [a2 , 0, c2 ] + [a1 , 0, c1 ]
=y+x

Property A3 is satisfied. Setting a = 0 and c = 0, we find that 0 = [0, 0, 0] ∈ V, so property A4 is


satisfied. Since −a1 and −c1 are members of R, −x ∈ V, and property A5 is satisfied.
Choose α1 ∈ R and β1 ∈ R.

α1 x = α1 [a1 , 0, c1 ]
= [α1 a1 , 0, α1 c1 ]

Since α1 a1 and α1 c1 are members of R, α1 x ∈ V, and property M1 is satisfied.

(α1 β1 )x = (α1 β1 )[a1 , 0, c1 ]


= [α1 β1 a1 , 0, α1 β1 c1 ]
= α1 [β1 a1 , 0, β1 c1 ]
= α1 (β1 x)

So property M2 is satisfied.

α1 (x + y) = α1 ([a1 , 0, c1 ] + [a2 , 0, c2 ])
= α1 [a1 + a2 , 0, c1 + c2 ]
= [α1 (a1 + a2 ), 0, α1 (c1 + c2 )]
= [α1 a1 + α1 a2 , 0, α1 c1 + α1 c2 ]
= [α1 a1 , 0, α1 c1 ] + [α1 a2 , 0, α1 c2 ]
= α1 [a1 , 0, c1 ] + α1 [a2 , 0, c2 ]
= α1 x + α1 y

So property M3 is satisfied.

(α1 + β1 )x = (α1 + β1 )[a1 , 0, c1 ]


= [(α1 + β1 )a1 , 0, (α1 + β1 )c1 ]
= [α1 a1 + β1 a1 , 0, α1 c1 + β1 c1 ]
= [α1 a1 , 0, α1 c1 ] + [β1 a1 , 0, β1 c1 ]
= α1 [a1 , 0, c1 ] + β1 [a1 , 0, c1 ]
= α1 x + β1 x

So property M4 is satisfied. 1x = 1[a1 , 0, c1 ] = [1 × a1 , 1 × 0, 1 × c1 ] = [a1 , 0, c1 ] = x, so property


M5 is satisfied.

All ten properties are satisfied, so V = {[a, 0, c] | a, c ∈ R} is a vector space.

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 5 Page 3 of 6

Part (b)

V = {[a, 1, c] | a, c ∈ R}
There is no 0 = [0, 0, 0] vector in this set of 3-vectors, so property A4 is not satisfied. Therefore,
V = {[a, 1, c] | a, c ∈ R} is not a vector space.

Part (c)

V = {[a, b, c] | ab = 0 and a, b, c ∈ R}
Choose x = [a1 , b1 , c1 ] ∈ V and y = [a2 , b2 , c2 ] ∈ V and z = [a3 , b3 , c3 ] ∈ V.

x + y = [a1 , b1 , c1 ] + [a2 , b2 , c2 ] = [a1 + a2 , b1 + b2 , c1 + c2 ]

The conditions a1 b1 = 0 and a2 b2 = 0 do not guarantee that (a1 + a2 )(b1 + b2 ) = 0; for example,
take a1 = 0, b1 = 1, a2 = 1, and b2 = 0. Therefore, x + y 6∈ V, and
V = {[a, b, c] | ab = 0 and a, b, c ∈ R} is not a vector space.

Part (d)

V = {[a, b, c] | m[1, 1, 0] + n[2, 0, 1] = [a, b, c] = [m + 2n, m, n], m, n ∈ R}


Choose x = [m1 + 2n1 , m1 , n1 ] ∈ V and y = [m2 + 2n2 , m2 , n2 ] ∈ V and z = [m3 + 2n3 , m3 , n3 ] ∈ V.

x + y = [m1 + 2n1 , m1 , n1 ] + [m2 + 2n2 , m2 , n2 ]


= [m1 + 2n1 + m2 + 2n2 , m1 + m2 , n1 + n2 ]
= [(m1 + m2 ) + 2(n1 + n2 ), m1 + m2 , n1 + n2 ]
| {z } | {z }
m4 n4
= [m4 + 2n4 , m4 , n4 ]

Because m4 ∈ R and n4 ∈ R, x + y ∈ V. Hence, property A1 is satisfied.

(x + y) + z = ([m1 + 2n1 , m1 , n1 ] + [m2 + 2n2 , m2 , n2 ]) + [m3 + 2n3 , m3 , n3 ]


= [m1 + 2n1 + m2 + 2n2 , m1 + m2 , n1 + n2 ] + [m3 + 2n3 , m3 , n3 ]
= [m1 + 2n1 + m2 + 2n2 + m3 + 2n3 , m1 + m2 + m3 , n1 + n2 + n3 ]
= [m1 + 2n1 , m1 , n1 ] + [m2 + 2n2 + m3 + 2n3 , m2 + m3 , n2 + n3 ]
= [m1 + 2n1 , m1 , n1 ] + ([m2 + 2n2 , m2 , n2 ] + [m3 + 2n3 , m3 , n3 ])
= x + (y + z)

Property A2 is satisfied.

x + y = [m1 + 2n1 , m1 , n1 ] + [m2 + 2n2 , m2 , n2 ]


= [m1 + 2n1 + m2 + 2n2 , m1 + m2 , n1 + n2 ]
= [m2 + 2n2 + m1 + 2n1 , m2 + m1 , n2 + n1 ]
= [m2 + 2n2 , m2 , n2 ] + [m1 + 2n1 , m1 , n1 ]
=y+x

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 5 Page 4 of 6

Property A3 is satisfied. Setting m = 0 and n = 0, we find that 0 = [0, 0, 0] ∈ V, so property A4 is


satisfied. Since −m1 and −n1 and −(m1 + 2n1 ) are members of R, −x ∈ V, and property A5 is
satisfied.
Choose α1 ∈ R and β1 ∈ R.

α1 x = α1 [m1 + 2n1 , m1 , n1 ]
= [α1 (m1 + 2n1 ), α1 m1 , α1 n1 ]

Since α1 (m1 + 2n1 ) and α1 m1 and α1 n1 are members of R, α1 x ∈ V, and property M1 is satisfied.

(α1 β1 )x = (α1 β1 )[m1 + 2n1 , m1 , n1 ]


= [α1 β1 (m1 + 2n1 ), α1 β1 m1 , α1 β1 n1 ]
= α1 [β1 (m1 + 2n1 ), β1 m1 , β1 n1 ]
= α1 (β1 x)

So property M2 is satisfied.

α1 (x + y) = α1 ([m1 + 2n1 , m1 , n1 ] + [m2 + 2n2 , m2 , n2 ])


= α1 [m1 + 2n1 + m2 + 2n2 , m1 + m2 , n1 + n2 ]
= [α1 (m1 + 2n1 + m2 + 2n2 ), α1 (m1 + m2 ), α1 (n1 + n2 )]
= [α1 (m1 + 2n1 ) + α1 (m2 + 2n2 ), α1 m1 + α1 m2 , α1 n1 + α1 n2 ]
= [α1 (m1 + 2n1 ), α1 m1 , α1 n1 ] + [α1 (m2 + 2n2 ), α1 m2 , α1 n2 ]
= α1 [m1 + 2n1 , m1 , n1 ] + α1 [m2 + 2n2 , m2 , n2 ]
= α1 x + α1 y

So property M3 is satisfied.

(α1 + β1 )x = (α1 + β1 )[m1 + 2n1 , m1 , n1 ]


= [(α1 + β1 )(m1 + 2n1 ), (α1 + β1 )m1 , (α1 + β1 )n1 ]
= [α1 (m1 + 2n1 ) + β1 (m1 + 2n1 ), α1 m1 + β1 m1 , α1 n1 + β1 n1 ]
= [α1 (m1 + 2n1 ), α1 m1 , α1 n1 ] + [β1 (m1 + 2n1 ), β1 m1 , β1 n1 ]
= α1 [m1 + 2n1 , m1 , n1 ] + β1 [m1 + 2n1 , m1 , n1 ]
= α1 x + β 1 x

So property M4 is satisfied.
1x = 1[m1 + 2n1 , m1 , n1 ] = [1(m1 + 2n1 ), 1 × m1 , 1 × n1 ] = [m1 + 2n1 , m1 , n1 ] = x, so property M5
is satisfied.

All ten properties are satisfied, so V = {[a, b, c] | m[1, 1, 0] + n[2, 0, 1] = [a, b, c]} is a vector space.

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 5 Page 5 of 6

Part (e)

V = {[a, b, c] | c − a = 2b, a, c ∈ R}
Choose x = [a1 , 21 (c1 − a1 ), c1 ] ∈ V and y = [a2 , 12 (c2 − a2 ), c2 ] ∈ V and z = [a3 , 21 (c3 − a3 ), c3 ] ∈ V.
 
   
1 1 1
x + y = a1 , (c1 − a1 ), c1 + a2 , (c2 − a2 ), c2 = a1 + a2 , {(c1 + c2 ) − (a1 + a2 )}, c1 + c2 
2 2 | {z } 2 | {z }
a4 c4
 
1
= a4 , (c4 − a4 ), c4
2

Because a4 ∈ R and c4 ∈ R and the second component is 21 (c4 − a4 ), x + y ∈ V. Hence, property


A1 is satisfied.
     
1 1 1
(x + y) + z = a1 , (c1 − a1 ), c1 + a2 , (c2 − a2 ), c2 + a3 , (c3 − a3 ), c3
2 2 2
   
1 1
= a1 + a2 , {(c1 + c2 ) − (a1 + a2 )}, c1 + c2 + a3 , (c3 − a3 ), c3
2 2
 
1
= a1 + a2 + a3 , {(c1 + c2 + c3 ) − (a1 + a2 + a3 )}, c1 + c2 + c3
2
   
1 1
= a1 , (c1 − a1 ), c1 + a2 + a3 , {(c2 + c3 ) − (a2 + a3 ), c2 + c3
2 2
     
1 1 1
= a1 , (c1 − a1 ), c1 + a2 , (c2 − a2 ), c2 + a3 , (c3 − a3 ), c3
2 2 2
= x + (y + z)

Property A2 is satisfied.
   
1 1
x + y = a1 , (c1 − a1 ), c1 + a2 , (c2 − a2 ), c2
2 2
 
1
= a1 + a2 , {(c1 + c2 ) − (a1 + a2 )}, c1 + c2
2
 
1
= a2 + a1 , {(c2 + c1 ) − (a2 + a1 )}, c2 + c1
2
   
1 1
= a2 , (c2 − a2 ), c2 + a1 , (c1 − a1 ), c1
2 2
=y+x

Property A3 is satisfied. Setting a = 0 and c = 0, we find that 0 = [0, 0, 0] ∈ V, so property A4 is


satisfied. Since −a1 and −c1 are members of R, −x ∈ V, and property A5 is satisfied.
Choose α1 ∈ R and β1 ∈ R.
 
1
α1 x = α1 a1 , (c1 − a1 ), c1
2
 
1
= α1 a1 , α1 (c1 − a1 ), α1 c1
2

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 5 Page 6 of 6

Since α1 a1 and α1 c1 and 21 α1 (c1 − a1 ) are members of R, α1 x ∈ V, and property M1 is satisfied.


 
1
(α1 β1 )x = (α1 β1 ) a1 , (c1 − a1 ), c1
2
 
1
= α1 β1 a1 , α1 β1 (c1 − a1 ), α1 β1 c1
2
 
1
= α1 β1 a1 , β1 (c1 − a1 ), β1 c1
2
= α1 (β1 x)

So property M2 is satisfied.
   
1 1
α1 (x + y) = α1 a1 , (c1 − a1 ), c1 + a2 , (c2 − a2 ), c2
2 2
 
1
= α1 a1 + a2 , {(c1 + c2 ) − (a1 + a2 )}, c1 + c2
2
 
1
= α1 (a1 + a2 ), α1 {(c1 + c2 ) − (a1 + a2 )}, α1 (c1 + c2 )
2
 
1 1
= α1 a1 + α1 a2 , α1 (c1 − a1 ) + α1 (c2 − a2 ), α1 c1 + α1 c2
2 2
   
1 1
= α1 a1 , α1 (c1 − a1 ), α1 c1 + α1 a2 , α1 (c2 − a2 ), α1 c2
2 2
   
1 1
= α1 a1 , (c1 − a1 ), c1 + α1 a2 , (c2 − a2 ), c2
2 2
= α1 x + α1 y

So property M3 is satisfied.
 
1
(α1 + β1 )x = (α1 + β1 ) a1 , (c1 − a1 ), c1
2
 
1
= (α1 + β1 )a1 , (α1 + β1 )(c1 − a1 ), (α1 + β1 )c1
2
 
1 1
= α1 a1 + β1 a1 , α1 (c1 − a1 ) + β1 (c1 − a1 ), α1 c1 + β1 c1
2 2
 
1 1
= α1 a1 , α1 (c1 − a1 ), α1 c1 ] + [β1 a1 , β1 (c1 − a1 ), β1 c1
2 2
   
1 1
= α1 a1 , (c1 − a1 ), c1 + β1 a1 , (c1 − a1 ), c1
2 2
= α1 x + β1 x

So property M4 is satisfied.
1x = 1[a1 , 21 (c1 − a1 ), c1 ] = [1 × a1 , 1 × 12 (c1 − a1 ), 1 × c1 = [a1 , 21 (c1 − a1 ), c1 ] = x, so property M5
is satisfied. All ten properties are satisfied, so V = {[a, b, c] | c − a = 2b, a, c ∈ R} is a vector space.

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 6 Page 1 of 1

Exercise 6
Are the three vectors [1, 2, 3], [−2, 0, 1], and [1, 10, 17] linearly dependent or independent? Do they
span all vectors or not?

Solution

By definition, a set of vectors S = {v1 , v2 , . . . , vn } is said to be linearly independent when the


only solution to the homogeneous equation

α1 v1 + α2 v2 + · · · + αn vn = 0

is the trivial solution α1 = α2 = · · · = αn = 0. Otherwise, the set is linearly dependent.

For this particular exercise, v1 = [1, 2, 3] and v2 = [−2, 0, 1] and v3 = [1, 10, 17]. There is a
nontrivial solution:
5v1 + 2v2 + (−1)v3 = 0
Therefore, the three vectors are linearly dependent. A more systematic way of coming to this
conclusion is the following: Arrange the vectors as the columns (or rows) of a matrix and find the
determinant of this matrix. If the determinant is nonzero, then the vectors are linearly
independent; otherwise, they are linearly dependent.
 
1 2 3
0 1

−2 1

−2 0

det −2 0 1  = 1 − 2
1 17 + 3 1 10

10 17
1 10 17
= 1(−10) − 2(−34 − 1) + 3(−20)
=0

This confirms the result. Because v3 can be written in terms of v1 and v2 , the vectors only span
a plane in R3 , not all of R3 .

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 7 Page 1 of 1

Exercise 7
Are the functions 1 + x, 1 − x, and 1 + x + x2 linearly dependent or independent? Why?

Solution

To determine whether a set of functions S = {f1 (x), f2 (x), . . . , fn (x)} is linearly independent or
not, one must consider the Wronski matrix.

···
 
f1 (x) f2 (x) fn (x)
 f10 (x) f20 (x) ··· fn0 (x) 
W(x) =  .. .. ..
 
..
.

 . . . 
(n−1) (n−1) (n−1)
f1 (x) f2 (x) · · · fn (x)

If there is a point x = x0 such that W(x0 ) is nonsingular (i.e. the determinant is nonzero), then
the set of functions is linearly independent.

The associated Wronski matrix for the set of functions in this exercise is
1 + x 1 − x 1 + x + x2
 

W(x) =  1 −1 1 + 2x 
0 0 2

1 + x 1 − x 1 + x + x2


det W(x) = 1 −1 1 + 2x
0 0 2

−1 1 + 2x 1 1 + 2x 2 1 −1

= (1 + x)
− (1 − x)
+ (1 + x + x )
0 2 0 2 0 0
= (1 + x)(−2) − (1 − x)(2) + (1 + x + x2 )(0)
= −4

The determinant of W(x) is nonzero for all values of x, so W(x) is nonsingular. Therefore, the
set of functions {1 + x, 1 − x, 1 + x + x2 } is linearly independent.

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 8 Page 1 of 1

Exercise 8
Find a vector that, together with the vectors [1, 1, 1] and [1, 2, 1], forms a basis of R3 .

Solution

By definition, a basis for R3 is a set of linearly independent vectors that span R3 . In other words,
if every vector in R3 can be written as a linear combination of the vectors in a set, then this set is
a basis of R3 .

The vectors v1 = [1, 1, 1] and v2 = [1, 2, 1] point in two distinct directions in R3 , so it is necessary
for the third vector to point in a different direction. Thus, let the third vector be formed from the
cross product of the two, v3 = v1 × v2 . v3 will point in a direction perpendicular to v1 and v2 .

x̂ ŷ ẑ

v3 = 1 1 1
1 2 1

1 1 1 1 1 1
= x̂ + ŷ
1 1 + ẑ 1 2

2 1
= x̂(−1) + ŷ(0) + ẑ(1)

Therefore, the three vectors, [1, 1, 1], [1, 2, 1], and [−1, 0, 1], form a basis of R3 .

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 9 Page 1 of 2

Exercise 9
Show that the functions (c1 + c2 sin2 x + c3 cos2 x) form a vector space. Find a basis of it. What is
its dimension?

Solution

This is a linear combination of the functions 1, sin2 x, and cos2 x. Since sin2 x + cos2 x = 1, the
basis is just {sin2 x, cos2 x}. And because the basis has two components to it, the dimension is 2.
In order to show that the functions form a vector space, we must show that the following ten
properties are satisfied:

(A1) x + y ∈ V for all x, y ∈ V.

(A2) (x + y) + z = x + (y + z) for every x, y, z ∈ V.

(A3) x + y = y + x for every x, y ∈ V.

(A4) There is an element 0 ∈ V such that x + 0 = x for every x ∈ V.

(A5) For each x ∈ V, there is an element (−x) ∈ V such that x + (−x) = 0.

(M1) αx ∈ V for all α ∈ R and x ∈ V.

(M2) (αβ)x = α(βx) for all α, β ∈ R and every x ∈ V.

(M3) α(x + y) = αx + αy for every α ∈ R and all x, y ∈ V.

(M4) (α + β)x = αx + βx for all α, β ∈ R and every x ∈ V.

(M5) 1x = x for every x ∈ V.

Suppose x ∈ V and y ∈ V. Then

x = a1 + b1 sin2 x + c1 cos2 x
y = a2 + b2 sin2 x + c2 cos2 x

The sum is x + y.

x + y = a1 + b1 sin2 x + c1 cos2 x + a2 + b2 sin2 x + c2 cos2 x


= (a1 + a2 ) + (b1 + b2 ) sin2 x + (c1 + c2 ) cos2 x
= a + b sin2 x + c cos2 x

Because a, b, c ∈ R, x + y is just another vector in V; that is, x + y ∈ V, and property A1 is


satisfied. Suppose z ∈ V. Then z = a3 + b3 sin2 x + c3 cos2 x, and

(x + y) + z = (a1 + b1 sin2 x + c1 cos2 x + a2 + b2 sin2 x + c2 cos2 x) + a3 + b3 sin2 x + c3 cos2 x


= a1 + b1 sin2 x + c1 cos2 x + a2 + b2 sin2 x + c2 cos2 x + a3 + b3 sin2 x + c3 cos2 x
= a1 + b1 sin2 x + c1 cos2 x + (a2 + b2 sin2 x + c2 cos2 x + a3 + b3 sin2 x + c3 cos2 x)
= x + (y + z)

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 9 Page 2 of 2

So property A2 is satisfied.
x + y = a1 + b1 sin2 x + c1 cos2 x + a2 + b2 sin2 x + c2 cos2 x
= a2 + b2 sin2 x + c2 cos2 x + a1 + b1 sin2 x + c1 cos2 x
=y+x
So property A3 is satisfied. We can choose a1 = b1 = c1 = 0 to get the zero vector, so property A4
is satisfied. We can choose a2 = −a1 , b2 = −b1 , and c2 = −c1 in y to get −x.
x + (−x) = a1 + b1 sin2 x + c1 cos2 x + (−a1 − b1 sin2 x − c1 cos2 x)
XX2X X2X
X
x − XX2X X2X X
1 − b1 sin x −
X  X 
=
a
1 + b1 sin x +
c1X
cos
  a c1X
cos
  x
=0
So property A5 is satisfied. Choose α ∈ R. Then
αx = α(a1 + b1 sin2 x + c1 cos2 x)
= αa1 + αb1 sin2 x + αc1 cos2 x
= a + b sin2 x + c cos2 x
Thus, αx is just another vector in V. So property M1 is satisfied. Choose β ∈ R. Then
(αβ)x = (αβ)(a1 + b1 sin2 x + c1 cos2 x)
= αβa1 + αβb1 sin2 x + αβc1 cos2 x
= α(βa1 + βb1 sin2 x + βc1 cos2 x)
= α(βx)
So property M2 is satisfied.
α(x + y) = α(a1 + b1 sin2 x + c1 cos2 x + a2 + b2 sin2 x + c2 cos2 x)
= αa1 + αb1 sin2 x + αc1 cos2 x + αa2 + αb2 sin2 x + αc2 cos2 x
= α(a1 + b1 sin2 x + c1 cos2 x) + α(a2 + b2 sin2 x + c2 cos2 x)
= αx + αy
So property M3 is satisfied.
(α + β)x = (α + β)(a1 + b1 sin2 x + c1 cos2 x)
= (α + β)a1 + (α + β)b1 sin2 x + (α + β)c1 cos2 x
= αa1 + βa1 + αb1 sin2 x + βb1 sin2 x + αc1 cos2 x + βc1 cos2 x
= αa1 + αb1 sin2 x + αc1 cos2 x + βa1 + βb1 sin2 x + βc1 cos2 x
= α(a1 + b1 sin2 x + c1 cos2 x) + β(a2 + b2 sin2 x + c2 cos2 x)
= αx + βy
So property M4 is satisfied.
1x = 1(a1 + b1 sin2 x + c1 cos2 x)
= 1 ∗ a1 + 1 ∗ b1 sin2 x + 1 ∗ c1 cos2 x
= a1 + b1 sin2 x + c1 cos2 x
=x
So property M5 is satisfied. All ten properties are satisfied, and so the functions
(c1 + c2 sin2 x + c3 cos2 x) form a vector space.

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 10 Page 1 of 3

Exercise 10
Show that the solutions of the differential equation u000 − 3u00 + 4u = 0 form a vector space. Find a
basis of it.

Solution

In order for the solutions of a differential equation to form a vector space U, the vector addition
and scalar multiplication operations must satisfy the following ten properties:
(A1) u1 + u2 ∈ U for all u1 , u2 ∈ U.

(A2) (u1 + u2 ) + u3 = u1 + (u2 + u3 ) for every u1 , u2 , u3 ∈ U.

(A3) u1 + u2 = u2 + u1 for every u1 , u2 ∈ U.

(A4) There is an element 0 ∈ U such that u1 + 0 = u1 for every u1 ∈ U.

(A5) For each u1 ∈ U, there is an element (−u1 ) ∈ U such that u1 + (−u1 ) = 0.

(M1) αu1 ∈ U for all α ∈ R and u1 ∈ U.

(M2) (αβ)u1 = α(βu1 ) for all α, β ∈ R and every u1 ∈ U.

(M3) α(u1 + u2 ) = αu1 + αu2 for every α ∈ R and all u1 , u2 ∈ U.

(M4) (α + β)u1 = αu1 + βu1 for all α, β ∈ R and every u1 ∈ U.

(M5) 1u1 = u1 for every u1 ∈ U.


Choose u1 ∈ U and u2 ∈ U and u3 ∈ U. Then

u000 00
1 − 3u1 + 4u1 = 0
u000 00
2 − 3u2 + 4u2 = 0
u000 00
3 − 3u3 + 4u3 = 0

Adding the first two equations together, we get

u000 00 000 00
1 − 3u1 + 4u1 + u2 − 3u2 + 4u2 = 0 + 0
u000 000 00 00
1 + u2 − 3u1 − 3u2 + 4u1 + 4u2 = 0
(u1 + u2 )000 − 3(u1 + u2 )00 + 4(u1 + u2 ) = 0

So u1 + u2 also satisfies the differential equation; in other words, u1 + u2 ∈ U, and property A1 is


satisfied.

[(u1 + u2 ) + u3 ]000 − 3[(u1 + u2 ) + u3 ]00 + 4[(u1 + u2 ) + u3 ] = 0


[u1 + u2 + u3 ]000 − 3[u1 + u2 + u3 ]00 + 4[u1 + u2 + u3 ] = 0
[u1 + (u2 + u3 )]000 − 3[u1 + (u2 + u3 )]00 + 4[u1 + (u2 + u3 )] = 0

Therefore, property A2 is satisfied.

(u1 + u2 )000 − 3(u1 + u2 )00 + 4(u1 + u2 ) = 0


(u2 + u1 )000 − 3(u2 + u1 )00 + 4(u2 + u1 ) = 0

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 10 Page 2 of 3

Therefore, property A3 is satisfied. Setting u = 0 in the differential equation gives equality, so


0 ∈ U. Property A4 is satisfied.

u000 00
1 − 3u1 + 4u1 = 0
−u000 00
1 + 3u1 − 4u1 = 0
(−u1 )000 − 3(−u1 )00 + 4(−u1 ) = 0

Thus, −u1 is a solution of the differential equation, i.e. −u1 ∈ U, and property A5 is satisfied.

Choose α ∈ R and β ∈ R.

α(u000 00
1 − 3u1 + 4u1 ) = α × 0
αu000 00
1 − 3αu1 + 4αu1 = 0
(αu1 )000 − 3(αu1 )00 + 4(αu1 ) = 0

So αu1 also satisfies the differential equation; in other words, αu1 ∈ U, and property M1 is
satisfied.

(αβ)(u000 00
1 − 3u1 + 4u1 ) = (αβ) × 0
(αβ)u000 00
1 − 3(αβ)u1 + 4(αβ)u1 = 0
[(αβ)u1 ]000 − 3[(αβ)u1 ]00 + 4[(αβ)u1 ] = 0
[α(βu1 )]000 − 3[α(βu1 )]00 + 4[α(βu1 )] = 0

Both (αβ)u1 and α(βu1 ) satisfy the differential equation, so (αβ)u1 = α(βu1 ). And property M2
is satisfied.

[α(u1 + u2 )]000 − 3[α(u1 + u2 )]00 + 4[α(u1 + u2 )] = 0


α(u000 00 000 00
1 − 3u1 + 4u1 + u2 − 3u2 + 4u2 ) = 0
αu000 00 000 00
1 − 3αu1 + 4αu1 + αu2 − 3αu2 + 4αu2 = 0
α(u000 00 000 00
1 − 3u1 + 4u1 ) + α(u2 − 3u2 + 4u2 ) = 0

Thus, property M3 is satisfied.

(α + β)(u000 00
1 − 3u1 + 4u1 ) = (α + β) × 0
(α + β)u000 00
1 − 3(α + β)u1 + 4(α + β)u1 = 0
αu000 00 000 00
1 − 3αu1 + 4αu1 + βu1 − 3βu1 + 4βu1 = 0
α(u000 00 000 00
1 − 3u1 + 4u1 ) + β(u1 − 3u1 + 4u1 ) = 0

And so property M4 holds.

(1u1 )000 − 3(1u1 )00 + 4(1u1 ) = 0


1 ∗ u000 00
1 − 1 ∗ 3u1 + 1 ∗ 4u1 = 0
u000 00
1 − 3u1 + 4u1 = 0

And so property M5 is satisfied. All ten properties are satisfied, so the solutions of the differential
equation form a vector space.

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 10 Page 3 of 3

The equation u000 − 3u00 + 4u = 0 is a linear ordinary differential equation with constant
coefficients. Therefore, the solutions are of the form u(x) = erx . Plugging this into the equation,

r3 erx − 3r2 erx + 4erx = 0


r3 − 3r2 + 4 = 0
(r + 1)(r2 − 4r + 4) = 0
(r + 1)(r − 2)2 = 0
→ r = {−1, 2}

So the solution to the differential equation is

u(x) = c1 e−x + c2 e2x + c3 xe2x

The basis of the vector space formed by the solutions is therefore {e−x , e2x , xe2x }.

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 11 Page 1 of 1

Exercise 11
Verify that u(x, y) = f (x)g(y) is a solution of the PDE uuxy = ux uy for all pairs of (differentiable)
functions f and g of one variable.

Solution

Suppose that f and g are differentiable functions of x and y, respectively. Then

ux = f 0 (x)g(y)
uy = f (x)g 0 (y)
uxy = f 0 (x)g 0 (y)

Substituting the expressions,

uuxy = [f (x)g(y)][f 0 (x)g 0 (y)]


= f (x)g(y)f 0 (x)g 0 (y)
= f 0 (x)g(y)f (x)g 0 (y)
= [f 0 (x)g(y)][f (x)g 0 (y)]
= ux uy

Therefore, u(x, y) = f (x)g(y) is a solution of the PDE, provided that the functions f and g are
differentiable.

www.stemjock.com
Strauss PDEs 2e: Section 1.1 - Exercise 12 Page 1 of 1

Exercise 12
Verify by direct substitution that

un (x, y) = sin nx sinh ny

is a solution of uxx + uyy = 0 for every n > 0.

Solution

Suppose n is some number greater than zero. Then

ux = n cos nx sinh ny
uxx = −n2 sin nx sinh ny
uy = n sin nx cosh ny
uyy = n2 sin nx sinh ny

Substituting the expressions,

uxx + uyy = −n2 sin nx sinh ny + n2 sin nx sinh ny


= (−n2 + n2 ) sin nx sinh ny
= (0) sin nx sinh ny
=0

Therefore, u = un (x, y) = sin nx sinh ny is a solution of the PDE for every n > 0.

www.stemjock.com

You might also like