7 3 Eigenvalue Method For Linear Systems
7 3 Eigenvalue Method For Linear Systems
7 3 Eigenvalue Method For Linear Systems
xn a n1 x 1 a n2 x 2 a nn x n .
Proof: By theorem from previous section, it suffices to find n linearly independent solution vectors z 1 , , z n,
and place them in a linear combination. But how to find these vectors?
Recall the educated guess when solving individual homogeneous linear DEQs: x t e rt . But this isn’t a vector.
t
v1 v1e
t
t
v2 t
v2e
So how about z t ve e , for some constants , v i .
t
vn vne
Note that z v e t.
t
To be a solution, we need z A z , or substituting from above, v e A v e t,
and canceling e t from both sides gives us v Av.
You may recall that when v 0, then the above equation is the requirement for having an
eigenvalue, eigenvector pair , v of A.
Notationally manipulate z t into the form f t ig t to identify the real and imaginary parts (see examples of this below)
Once you have found these two solutions ( f t and g t ), you are done.
The solutions associated with the eigenvalue’s conjugate are identical.
As an exercise, verify that the solutions obtained from each conjugate , are identical.
See the examples below to get a better feel for what this section is saying.
Exercises
Problem: #12 Apply the eigenvalue method of this section to find a general solution of the given system:
x 1 x 1 5x 2 , x 2 x 1 3x 2 .
1 5
A
1 3
1 5 2
Characteristic Equation: |A I| 1 3 5 4 8.
1 3
4 16 32
2
Eigenvalues: 2 2i.
1 1 2i
, y b, x b1 2i
0 0
T T
Eigenvector: v b1 2i b 1 2i 1 , when b 1.
2 2i t
So we have: v e (from above, "Notationally manipulate into the form f t ig t ")
1 2i
e 2t e 2it
1
1 2i cos 2t i sin 2t 1 2i
e 2t cos 2t i sin 2t e 2t .
1 cos 2t i sin 2t
The image below shows a direction field for this DEQ and some typical solution curves:
Problem: #25 Apply the eigenvalue method to find a general solution of the system.
x 1 5x 1 5x 2 2x 3 , x2 6x 1 6x 2 5x 3 , x 3 6x 1 6x 2 5x 3
5 5 2
A 6 6 5
6 6 5
3 2
Characteristic equation: 4 13 0
Eigenvalues: 0 and 2 3i
T
So: z 1 t v 1e0 t 1 1 0 .
1 1
6 6 3 3i 1 1 2 2
i
6 8 3i 5 0 2 3i 2 3i
3 3i 5 2 3 3i 5 2
1 1 3 3 3 3
Note: 3 3i 2 2
i 2 2 2
i 2
i 3i. So:
1 1
1 1 2 2
i
0 2 3i 2 3i
0 2 3i 2 3i
1 1 1 1
1 1 2 2
i 1 0 2 2
i
0 1 1 0 1 1
0 0 0 0 0 0
c c
z c, y c, x 2 2
i.
c c
2 2
i 1 1i
c 2 where c 2.
c 2
T
Complex valued eigenvector: v 2 1 i 2 2 .
Problem: #34
This problem deals with the open three tank system. Freshwater flows into tank-1.
Mixed brine (salt water) flows from tank-1 into tank-2, from tank-2 into tank-3, and out of tank-3.
All have the flow rate r 60 gallons per minute. Initial (t 0) amounts of salt are:
x1 0 40 lb, x 2 0 0, and x 3 0 0 in the three tanks.
Initial volumes: V 1 20, V2 12, V3 60.
a.) First, solve for the amounts of salt in the three tanks at time t.
x1
So, x 1 0 60 20
60 3x 1 .
3 0 0 x1
3 5 0 x2 Az.
0 5 1 x3
3 0 0
The coefficient matrix A 3 5 0 has as eigenvalues, its diagonal elements:
0 5 1
3t 5t
So: z v 1e v 2e v 3e t.
Or written as a system:
3t
x1 t 4c 1 e
3t 5t
x2 t 6c 1 e 4c 2 e
3t 5t
x3 t 15c 1 e 5c 2 e c3e t. Now What?
The initial conditions x 1 0 40, x 2 0 0, and x 3 0 0 give us c 1 10, c 2 15, c 3 75. So we have:
3t
x1 t 40e
3t 5t
x2 t 60e 60e
3t 5t
x3 t 150e 75e 75e t .
b.) Now, determine the maximal amount of salt that tank-3 ever contains.
Remember from calculus that you can find the local maximums and minimums by taking the derivative of the function, and
setting it equal to zero. For tank-3:
3t 5t t
x3 t 450e 375e 75e 0
1
Multiplying by nonzero 75e t :
4t 2t
5e 6e 1 0 Factoring this is the (not so) hard part.
2t 2t
5e 1 e 1 0
2t 2t
And observe that for the second factor: e 1 when ln e ln 1 ,
or equivalently when 2t 0, or t 0.
2t 1 2t 1
Now looking at the first factor, e 5
when ln e ln 5
,
1
or when 2t ln 5, or t 2
ln 5 0. 8 min 48 sec.
3t 5t 1
Since x 3 t 150e 75e 75e t , the maximum amount of salt ever in tank-3 is x 3 2
ln 5 21. 5 pounds,