Documente Academic
Documente Profesional
Documente Cultură
d
y1 (t) = −5y1 (t) + 2y2 (t) (1)
dt
d
y2 (t) = 6y1 (t) − 6y2 (t) (2)
dt
with initial conditions y1 (0) = 7 and y2 (0) = 7. Or equivalently in matrix form as:
" #
−5 2
Ay =
6 −6
R1 R2
I1 I2
+ Vin C1 C2
−
Figure 1: Two dimensional system: a circuit with two capacitors, like the one in lecture.
© UCB EECS 16B, Fall 2019. All Rights Reserved. This may not be publicly shared without explicit permission. 1
2B: Changing Coordinates @ 2019-09-17 00:12:54-07:00 2
The reason we spent the time in the previous discussion to find the eigenvalues and eigenspaces is because
we can use them as a convenient basis for coordinate transformation! Today we’ll solve this differential
equation in a simple way with this methodology.
(a) Change coordinates into the eigenbasis to re-express the differential equations in terms of new variables
zλ1 (t), zλ2 (t). (These variables represent eigenbasis-aligned coordinates.)
Answer:
~y =~vλ1 zλ1 +~vλ2 zλ2
" #" #
−1 2 zλ1
~y =
2 3 zλ2
We can define the change-of-coordinates matrix from the eigenbasis to our original basis as
" #
−1 2
V=
2 3
That is:
" # " #" #
d
dt zλ1 (t) −9 0 zλ1 (t)
d =
dt zλ2 (t)
0 −2 zλ2 (t)
(b) Solve the differential equation for zλi (t) in the eigenbasis.
Answer: First we get the initial condition:
" #" # " #
− 37 2
7 7 −1
~zλ (0) = 2 1 =
7 7 7 3
Then we solve based on the form of the problem and our previous differential equation experience:
" #
K1 e−9t
~zλ (t) =
K2 e−2t
© UCB EECS 16B, Fall 2019. All Rights Reserved. This may not be publicly shared without explicit permission. 2
2B: Changing Coordinates @ 2019-09-17 00:12:54-07:00 3
(c) Convert your solution back into the original coordinates to find yi (t).
Answer: " #" # " #
−1 2 −e−9t e−9t + 6e−2t
~y(t) = V~zλ (t) = =
2 3 3e−2t −2e−9t + 9e−2t
(d) We can solve this equation using a slightly shorter approach by observing that the solutions for yi (t)
will all be of the form
yi (t) = ∑ Ki,k eλk t
k
where λk is an eigenvalue of our differential equation relation matrix and the Ki,k are constants derived
from our initial conditions and the coordinate changes involved.
Since we have observed that the solutions will include eλit terms, once we have found the eigenvalues
for our differential equation matrix, we can guess the forms of the yi (t) as
" # " #
y1 (t) αeλ1t + β eλ2t
=
y2 (t) γeλ1t + κeλ2t
where α, β , γ, κ are all constants.
Take the derivative to write out " #
d
dt y1 (t) .
d
dt y2 (t)
Equating terms:
−9α = −5α + 2γ
−2β = −5β + 2κ
−9γ = 6α − 6γ
−2κ = 6β − 6κ
© UCB EECS 16B, Fall 2019. All Rights Reserved. This may not be publicly shared without explicit permission. 3
2B: Changing Coordinates @ 2019-09-17 00:12:54-07:00 4
This gives:
γ = −2α
3
κ= β
2
Rewriting:
" # " #
y1 (t) αe−9t + β e−2t
=
y2 (t) −αe−9t + 2β e−2t
we get:
α =1
β =6
" # " #
y1 (t) e−9t + 6e−2t
=
y2 (t) −2e−9t + 9e−2t
Contributors:
• Anant Sahai.
• Regina Eckert.
• Nathan Lambert.
• Pavan Bhargava.
© UCB EECS 16B, Fall 2019. All Rights Reserved. This may not be publicly shared without explicit permission. 4