2d homogeneous linear systems with constant...
TRANSCRIPT
2D Homogeneous Linear Systems withConstant Coefficients
— repeated eigenvalues
Xu-Yan Chen
Systems of Diff Eqs:d~x
dt= A~x
where ~x(t) =
[x1(t)x2(t)
], A is a 2× 2 real constant matrix
Things to explore:
I General solutions
I Initial value problems
I Geometric figures
I Solutions graphs x1 vs t & x2 vs tI Direction fields in the (x1, x2) planeI Phase portraits in the (x1, x2) plane
I Stability/instability of equilibrium (x1, x2) = (0, 0)
2D Systems:d~x
dt= A~x
What if A has repeated eigenvalues?Assume that the eigenvalues of A are: λ1 = λ2.
• Easy Cases: A =
[λ1 00 λ1
];
• Hard Cases: A 6=[λ1 00 λ1
], but λ1 = λ2.
Find Solutions in the Easy Cases: A = λ1I
I All vector ~x ∈ R2 satisfy (A− λ1I)~x = 0.The eigenspace of λ1 is the entire plane.
We can pick ~u1 =
[10
], ~u2 =
[01
]as linearly indep eigenvectors.
I General solutions are ~x(t) = C1eλ1t~u1 + C2e
λ1t~u2
⇒ ~x(t) = C1eλ1t
[10
]+ C2e
λ1t
[01
]⇒ ~x(t) = eλ1t
[C1
C2
]
2D Systems:d~x
dt= A~x
What if A has repeated eigenvalues?Assume that the eigenvalues of A are: λ1 = λ2.
• Easy Cases: A =
[λ1 00 λ1
];
• Hard Cases: A 6=[λ1 00 λ1
], but λ1 = λ2.
Find Solutions in the Easy Cases: A = λ1I
I All vector ~x ∈ R2 satisfy (A− λ1I)~x = 0.The eigenspace of λ1 is the entire plane.
We can pick ~u1 =
[10
], ~u2 =
[01
]as linearly indep eigenvectors.
I General solutions are ~x(t) = C1eλ1t~u1 + C2e
λ1t~u2
⇒ ~x(t) = C1eλ1t
[10
]+ C2e
λ1t
[01
]⇒ ~x(t) = eλ1t
[C1
C2
]
2D Systems:d~x
dt= A~x
What if A has repeated eigenvalues?Assume that the eigenvalues of A are: λ1 = λ2.
• Easy Cases: A =
[λ1 00 λ1
];
• Hard Cases: A 6=[λ1 00 λ1
], but λ1 = λ2.
Find Solutions in the Easy Cases: A = λ1I
I All vector ~x ∈ R2 satisfy (A− λ1I)~x = 0.The eigenspace of λ1 is the entire plane.
We can pick ~u1 =
[10
], ~u2 =
[01
]as linearly indep eigenvectors.
I General solutions are ~x(t) = C1eλ1t~u1 + C2e
λ1t~u2
⇒ ~x(t) = C1eλ1t
[10
]+ C2e
λ1t
[01
]⇒ ~x(t) = eλ1t
[C1
C2
]
2D Systems:d~x
dt= A~x
What if A has repeated eigenvalues?Assume that the eigenvalues of A are: λ1 = λ2.
• Easy Cases: A =
[λ1 00 λ1
];
• Hard Cases: A 6=[λ1 00 λ1
], but λ1 = λ2.
Find Solutions in the Easy Cases: A = λ1I
I All vector ~x ∈ R2 satisfy (A− λ1I)~x = 0.The eigenspace of λ1 is the entire plane.
We can pick ~u1 =
[10
], ~u2 =
[01
]as linearly indep eigenvectors.
I General solutions are ~x(t) = C1eλ1t~u1 + C2e
λ1t~u2
⇒ ~x(t) = C1eλ1t
[10
]+ C2e
λ1t
[01
]⇒ ~x(t) = eλ1t
[C1
C2
]
2D Systems:d~x
dt= A~x
What if A has repeated eigenvalues?Assume that the eigenvalues of A are: λ1 = λ2.
• Easy Cases: A =
[λ1 00 λ1
];
• Hard Cases: A 6=[λ1 00 λ1
], but λ1 = λ2.
Find Solutions in the Easy Cases: A = λ1I
I All vector ~x ∈ R2 satisfy (A− λ1I)~x = 0.The eigenspace of λ1 is the entire plane.
We can pick ~u1 =
[10
], ~u2 =
[01
]as linearly indep eigenvectors.
I General solutions are ~x(t) = C1eλ1t~u1 + C2e
λ1t~u2
⇒ ~x(t) = C1eλ1t
[10
]+ C2e
λ1t
[01
]⇒ ~x(t) = eλ1t
[C1
C2
]
2D Systems:d~x
dt= A~x
What if A has repeated eigenvalues?Assume that the eigenvalues of A are: λ1 = λ2.
• Easy Cases: A =
[λ1 00 λ1
];
• Hard Cases: A 6=[λ1 00 λ1
], but λ1 = λ2.
Find Solutions in the Easy Cases: A = λ1I
I All vector ~x ∈ R2 satisfy (A− λ1I)~x = 0.The eigenspace of λ1 is the entire plane.
We can pick ~u1 =
[10
], ~u2 =
[01
]as linearly indep eigenvectors.
I General solutions are ~x(t) = C1eλ1t~u1 + C2e
λ1t~u2
⇒ ~x(t) = C1eλ1t
[10
]+ C2e
λ1t
[01
]
⇒ ~x(t) = eλ1t
[C1
C2
]
2D Systems:d~x
dt= A~x
What if A has repeated eigenvalues?Assume that the eigenvalues of A are: λ1 = λ2.
• Easy Cases: A =
[λ1 00 λ1
];
• Hard Cases: A 6=[λ1 00 λ1
], but λ1 = λ2.
Find Solutions in the Easy Cases: A = λ1I
I All vector ~x ∈ R2 satisfy (A− λ1I)~x = 0.The eigenspace of λ1 is the entire plane.
We can pick ~u1 =
[10
], ~u2 =
[01
]as linearly indep eigenvectors.
I General solutions are ~x(t) = C1eλ1t~u1 + C2e
λ1t~u2
⇒ ~x(t) = C1eλ1t
[10
]+ C2e
λ1t
[01
]⇒ ~x(t) = eλ1t
[C1
C2
]
2D Systems:d~x
dt= A~x
Find Solutions in the Hard Cases: λ1 = λ2, but A 6= λ1I
I Find an eigenvector ~u for λ1, by solving
(A− λ1I)~x = 0.
Since A 6= λ1I, we can only pick one linearly indep eigenvector ~u.This gives partial solutions: ~x(t) = Ceλ1t~u.
Need more to get complete solution formula.
I Find a generalized eigenvector ~v by solving
(A− λ1I)~x = ~u.
Solutions of (A− λ1I)~x = ~u satisfy (A− λ1I)2~x = (A− λ1I)~u = 0.
Nonzero solutions of (A− λ1I)2~x = 0 are called generalized eigenvectors.
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u)
2D Systems:d~x
dt= A~x
Find Solutions in the Hard Cases: λ1 = λ2, but A 6= λ1I
I Find an eigenvector ~u for λ1, by solving
(A− λ1I)~x = 0.
Since A 6= λ1I, we can only pick one linearly indep eigenvector ~u.This gives partial solutions: ~x(t) = Ceλ1t~u.
Need more to get complete solution formula.
I Find a generalized eigenvector ~v by solving
(A− λ1I)~x = ~u.
Solutions of (A− λ1I)~x = ~u satisfy (A− λ1I)2~x = (A− λ1I)~u = 0.
Nonzero solutions of (A− λ1I)2~x = 0 are called generalized eigenvectors.
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u)
2D Systems:d~x
dt= A~x
Find Solutions in the Hard Cases: λ1 = λ2, but A 6= λ1I
I Find an eigenvector ~u for λ1, by solving
(A− λ1I)~x = 0.Since A 6= λ1I, we can only pick one linearly indep eigenvector ~u.This gives partial solutions: ~x(t) = Ceλ1t~u.
Need more to get complete solution formula.
I Find a generalized eigenvector ~v by solving
(A− λ1I)~x = ~u.
Solutions of (A− λ1I)~x = ~u satisfy (A− λ1I)2~x = (A− λ1I)~u = 0.
Nonzero solutions of (A− λ1I)2~x = 0 are called generalized eigenvectors.
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u)
2D Systems:d~x
dt= A~x
Find Solutions in the Hard Cases: λ1 = λ2, but A 6= λ1I
I Find an eigenvector ~u for λ1, by solving
(A− λ1I)~x = 0.Since A 6= λ1I, we can only pick one linearly indep eigenvector ~u.This gives partial solutions: ~x(t) = Ceλ1t~u.
Need more to get complete solution formula.
I Find a generalized eigenvector ~v by solving
(A− λ1I)~x = ~u.
Solutions of (A− λ1I)~x = ~u satisfy (A− λ1I)2~x = (A− λ1I)~u = 0.
Nonzero solutions of (A− λ1I)2~x = 0 are called generalized eigenvectors.
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u)
2D Systems:d~x
dt= A~x
Find Solutions in the Hard Cases: λ1 = λ2, but A 6= λ1I
I Find an eigenvector ~u for λ1, by solving
(A− λ1I)~x = 0.Since A 6= λ1I, we can only pick one linearly indep eigenvector ~u.This gives partial solutions: ~x(t) = Ceλ1t~u.
Need more to get complete solution formula.
I Find a generalized eigenvector ~v by solving
(A− λ1I)~x = ~u.
Solutions of (A− λ1I)~x = ~u satisfy (A− λ1I)2~x = (A− λ1I)~u = 0.
Nonzero solutions of (A− λ1I)2~x = 0 are called generalized eigenvectors.
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u)
2D Systems:d~x
dt= A~x
Find Solutions in the Hard Cases: λ1 = λ2, but A 6= λ1I
I Find an eigenvector ~u for λ1, by solving
(A− λ1I)~x = 0.Since A 6= λ1I, we can only pick one linearly indep eigenvector ~u.This gives partial solutions: ~x(t) = Ceλ1t~u.
Need more to get complete solution formula.
I Find a generalized eigenvector ~v by solving
(A− λ1I)~x = ~u.
Solutions of (A− λ1I)~x = ~u satisfy (A− λ1I)2~x = (A− λ1I)~u = 0.
Nonzero solutions of (A− λ1I)2~x = 0 are called generalized eigenvectors.
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u)
2D Systems ~x′ = A~x: phase portraits & stability
λ1 = λ2 < 0, A = λ1I :Attractive proper node,asymtotically stable
λ1 = λ2 < 0, A 6= λ1I :Attractive degenerate node,asymtotically stable
λ1 = λ2 > 0, A = λ1I :Repulsive proper node,unstable
λ1 = λ2 > 0, A 6= λ1I :Repulsive degenerate node,unstable
λ1 = λ2 = 0, A = 0 :Every point is a stable equilib-rium, but not asymp stable
λ1 = λ2 = 0, A 6= 0 :Laminated flow,unstable
Example 1. (Attractive Proper node)
Consider ~x′ = A~x, where A =
[−2 00 −2
].
(a) Find general solutions of ~x′ =
[−2 00 −2
]~x.
(b) Solve the initial value problem ~x′ =
[−2 00 −2
]~x, ~x(0) =
[23
].
(c) Sketch the phase portrait.
(d) Is the equilibrium (0, 0) stable, asymptotically stable, or unstable?
Example 1.d~x
dt=
[−2 00 −2
]~x
(a) Find general solutions.
I Eigenvalues of A are λ1 = λ2 = −2.
I (A− λ1I)~x = 0⇔ 0~x = 0: All ~x ∈ R2 are eigenvectors.
I General solutions are
~x(t) = C1e−2t
[10
]+ C2e
−2t[01
]⇔ ~x(t) = e−2t
[C1
C2
]
(b) Solved~x
dt=
[−2 00 −2
]~x, ~x(0) =
[23
].
I Use the initial condition:
~x(0) =
[23
]⇒
[C1
C2
]=
[23
]I The solution to the initial value problem:
~x(t) = e−2t[23
]
Example 1.d~x
dt=
[−2 00 −2
]~x
(a) Find general solutions.
I Eigenvalues of A are λ1 = λ2 = −2.
I (A− λ1I)~x = 0⇔ 0~x = 0: All ~x ∈ R2 are eigenvectors.
I General solutions are
~x(t) = C1e−2t
[10
]+ C2e
−2t[01
]⇔ ~x(t) = e−2t
[C1
C2
]
(b) Solved~x
dt=
[−2 00 −2
]~x, ~x(0) =
[23
].
I Use the initial condition:
~x(0) =
[23
]⇒
[C1
C2
]=
[23
]I The solution to the initial value problem:
~x(t) = e−2t[23
]
Example 1.d~x
dt=
[−2 00 −2
]~x
(a) Find general solutions.
I Eigenvalues of A are λ1 = λ2 = −2.
I (A− λ1I)~x = 0⇔ 0~x = 0: All ~x ∈ R2 are eigenvectors.
I General solutions are
~x(t) = C1e−2t
[10
]+ C2e
−2t[01
]⇔ ~x(t) = e−2t
[C1
C2
]
(b) Solved~x
dt=
[−2 00 −2
]~x, ~x(0) =
[23
].
I Use the initial condition:
~x(0) =
[23
]⇒
[C1
C2
]=
[23
]I The solution to the initial value problem:
~x(t) = e−2t[23
]
Example 1.d~x
dt=
[−2 00 −2
]~x
(a) Find general solutions.
I Eigenvalues of A are λ1 = λ2 = −2.
I (A− λ1I)~x = 0⇔ 0~x = 0: All ~x ∈ R2 are eigenvectors.
I General solutions are
~x(t) = C1e−2t
[10
]+ C2e
−2t[01
]⇔ ~x(t) = e−2t
[C1
C2
]
(b) Solved~x
dt=
[−2 00 −2
]~x, ~x(0) =
[23
].
I Use the initial condition:
~x(0) =
[23
]⇒
[C1
C2
]=
[23
]I The solution to the initial value problem:
~x(t) = e−2t[23
]
Example 1.d~x
dt=
[−2 00 −2
]~x
(a) Find general solutions.
I Eigenvalues of A are λ1 = λ2 = −2.
I (A− λ1I)~x = 0⇔ 0~x = 0: All ~x ∈ R2 are eigenvectors.
I General solutions are
~x(t) = C1e−2t
[10
]+ C2e
−2t[01
]⇔ ~x(t) = e−2t
[C1
C2
]
(b) Solved~x
dt=
[−2 00 −2
]~x, ~x(0) =
[23
].
I Use the initial condition:
~x(0) =
[23
]⇒
[C1
C2
]=
[23
]
I The solution to the initial value problem:
~x(t) = e−2t[23
]
Example 1.d~x
dt=
[−2 00 −2
]~x
(a) Find general solutions.
I Eigenvalues of A are λ1 = λ2 = −2.
I (A− λ1I)~x = 0⇔ 0~x = 0: All ~x ∈ R2 are eigenvectors.
I General solutions are
~x(t) = C1e−2t
[10
]+ C2e
−2t[01
]⇔ ~x(t) = e−2t
[C1
C2
]
(b) Solved~x
dt=
[−2 00 −2
]~x, ~x(0) =
[23
].
I Use the initial condition:
~x(0) =
[23
]⇒
[C1
C2
]=
[23
]I The solution to the initial value problem:
~x(t) = e−2t[23
]
Example 1 (c) Phase portrait ofd~x
dt=
[−2 00 −2
]~x
General solutions: ~x(t) = e−2t[C1
C2
]
All solutions decay to 0 in the same exponential rate λ1 = −2.The trajectories are lines converging to the origin.
(d) Stability?
The equilibrium (0, 0) isasymptotically stable.
We have anattractive proper node,when A = λ1I and λ1 < 0.
Example 1 (c) Phase portrait ofd~x
dt=
[−2 00 −2
]~x
General solutions: ~x(t) = e−2t[C1
C2
]All solutions decay to 0 in the same exponential rate λ1 = −2.The trajectories are lines converging to the origin.
(d) Stability?
The equilibrium (0, 0) isasymptotically stable.
We have anattractive proper node,when A = λ1I and λ1 < 0.
Example 1 (c) Phase portrait ofd~x
dt=
[−2 00 −2
]~x
General solutions: ~x(t) = e−2t[C1
C2
]All solutions decay to 0 in the same exponential rate λ1 = −2.The trajectories are lines converging to the origin.
(d) Stability?
The equilibrium (0, 0) isasymptotically stable.
We have anattractive proper node,when A = λ1I and λ1 < 0.
Example 1 (c) Phase portrait ofd~x
dt=
[−2 00 −2
]~x
General solutions: ~x(t) = e−2t[C1
C2
]All solutions decay to 0 in the same exponential rate λ1 = −2.The trajectories are lines converging to the origin.
(d) Stability?
The equilibrium (0, 0) isasymptotically stable.
We have anattractive proper node,when A = λ1I and λ1 < 0.
Example 1 (c) Phase portrait ofd~x
dt=
[−2 00 −2
]~x
General solutions: ~x(t) = e−2t[C1
C2
]All solutions decay to 0 in the same exponential rate λ1 = −2.The trajectories are lines converging to the origin.
(d) Stability?
The equilibrium (0, 0) isasymptotically stable.
We have anattractive proper node,when A = λ1I and λ1 < 0.
Example 2. (Repulsive proper node)
~x′ = A~x, where A =
[3 00 3
].
(a) Find general solutions of ~x′ =
[3 00 3
]~x.
(b) Solve the initial value problem ~x′ =
[3 00 3
]~x, ~x(0) =
[23
].
(c) Sketch the phase portrait.
(d) Is the equilibrium (0, 0) stable, asymptotically stable, or unstable?
Example 2.d~x
dt=
[3 00 3
]~x
(a) Find general solutions.
General solutions are ~x(t) = e3t[C1
C2
]
(b) Solved~x
dt=
[3 00 3
]~x, ~x(0) =
[23
].
The solution to the initial value problem: ~x(t) = e3t[23
]
Example 2.d~x
dt=
[3 00 3
]~x
(a) Find general solutions.
General solutions are ~x(t) = e3t[C1
C2
]
(b) Solved~x
dt=
[3 00 3
]~x, ~x(0) =
[23
].
The solution to the initial value problem: ~x(t) = e3t[23
]
Example 2.d~x
dt=
[3 00 3
]~x
(a) Find general solutions.
General solutions are ~x(t) = e3t[C1
C2
]
(b) Solved~x
dt=
[3 00 3
]~x, ~x(0) =
[23
].
The solution to the initial value problem: ~x(t) = e3t[23
]
Example 2.d~x
dt=
[3 00 3
]~x
(a) Find general solutions.
General solutions are ~x(t) = e3t[C1
C2
]
(b) Solved~x
dt=
[3 00 3
]~x, ~x(0) =
[23
].
The solution to the initial value problem: ~x(t) = e3t[23
]
Example 2 (c) Phase portrait ofd~x
dt=
[3 00 3
]~x
General solutions: ~x(t) = e3t[C1
C2
]
All solutions grow in the same exponential rate λ1 = 3.The trajectories are lines emanating from the origin.
(d) Stability?
The equilibrium (0, 0) isunstable.
We have arepulsive proper node,when A = λ1I and λ1 > 0.
Example 2 (c) Phase portrait ofd~x
dt=
[3 00 3
]~x
General solutions: ~x(t) = e3t[C1
C2
]All solutions grow in the same exponential rate λ1 = 3.The trajectories are lines emanating from the origin.
(d) Stability?
The equilibrium (0, 0) isunstable.
We have arepulsive proper node,when A = λ1I and λ1 > 0.
Example 2 (c) Phase portrait ofd~x
dt=
[3 00 3
]~x
General solutions: ~x(t) = e3t[C1
C2
]All solutions grow in the same exponential rate λ1 = 3.The trajectories are lines emanating from the origin.
(d) Stability?
The equilibrium (0, 0) isunstable.
We have arepulsive proper node,when A = λ1I and λ1 > 0.
Example 2 (c) Phase portrait ofd~x
dt=
[3 00 3
]~x
General solutions: ~x(t) = e3t[C1
C2
]All solutions grow in the same exponential rate λ1 = 3.The trajectories are lines emanating from the origin.
(d) Stability?
The equilibrium (0, 0) isunstable.
We have arepulsive proper node,when A = λ1I and λ1 > 0.
Example 2 (c) Phase portrait ofd~x
dt=
[3 00 3
]~x
General solutions: ~x(t) = e3t[C1
C2
]All solutions grow in the same exponential rate λ1 = 3.The trajectories are lines emanating from the origin.
(d) Stability?
The equilibrium (0, 0) isunstable.
We have arepulsive proper node,when A = λ1I and λ1 > 0.
Example 3. (attractive degenerate node)
Consider ~x′ = A~x, where A =
[−7 8−2 1
].
(a) Find general solutions of ~x′ =
[−7 8−2 1
]~x.
(b) Solve the initial value problem ~x′ =
[−7 8−2 1
]~x, ~x(0) =
[23
].
(c) Sketch the phase portrait.
(d) Is the equilibrium (0, 0) stable, asymptotically stable, or unstable?
Example 3 (a) ~x′ =
[−7 8−2 1
]~x
I Eigenvalues of A, by solving det(A− λI) = 0:
det
[−7− λ 8−2 1− λ
]= λ2 + 6λ+ 9 = 0 ⇒ λ1 = λ2 = −3
I Eigenvectors of A for λ1 = λ2 = −3, by solving (A− λ1I)~x = 0:
(A+ 3I)~x = 0⇔[−4 8−2 4
] [x1x2
]=
[00
]⇔
[x1x2
]= x2
[21
]
• Can only pick one linear indep eigenvector ~u =
[21
].
• Partial solutions: ~x(t) = Ce−3t[21
].
• Need more to get complete solution formula.
Example 3 (a) ~x′ =
[−7 8−2 1
]~x
I Eigenvalues of A, by solving det(A− λI) = 0:
det
[−7− λ 8−2 1− λ
]= λ2 + 6λ+ 9 = 0 ⇒ λ1 = λ2 = −3
I Eigenvectors of A for λ1 = λ2 = −3, by solving (A− λ1I)~x = 0:
(A+ 3I)~x = 0⇔[−4 8−2 4
] [x1x2
]=
[00
]⇔
[x1x2
]= x2
[21
]
• Can only pick one linear indep eigenvector ~u =
[21
].
• Partial solutions: ~x(t) = Ce−3t[21
].
• Need more to get complete solution formula.
Example 3 (a) ~x′ =
[−7 8−2 1
]~x
I Eigenvalues of A, by solving det(A− λI) = 0:
det
[−7− λ 8−2 1− λ
]= λ2 + 6λ+ 9 = 0 ⇒ λ1 = λ2 = −3
I Eigenvectors of A for λ1 = λ2 = −3, by solving (A− λ1I)~x = 0:
(A+ 3I)~x = 0⇔[−4 8−2 4
] [x1x2
]=
[00
]⇔
[x1x2
]= x2
[21
]
• Can only pick one linear indep eigenvector ~u =
[21
].
• Partial solutions: ~x(t) = Ce−3t[21
].
• Need more to get complete solution formula.
Example 3 (a) ~x′ =
[−7 8−2 1
]~x
I Eigenvalues of A, by solving det(A− λI) = 0:
det
[−7− λ 8−2 1− λ
]= λ2 + 6λ+ 9 = 0 ⇒ λ1 = λ2 = −3
I Eigenvectors of A for λ1 = λ2 = −3, by solving (A− λ1I)~x = 0:
(A+ 3I)~x = 0⇔[−4 8−2 4
] [x1x2
]=
[00
]⇔
[x1x2
]= x2
[21
]• Can only pick one linear indep eigenvector ~u =
[21
].
• Partial solutions: ~x(t) = Ce−3t[21
].
• Need more to get complete solution formula.
Example 3 (a)d~x
dt=
[−7 8−2 1
]~x
I Eigenvalues of A: λ1 = λ2 = −3
I An eigenvector for λ1 = λ2 = −3: ~u =
[21
]
I Find a generalized eigenvector, by solving (A− λ1I)~x = ~u:
(A+ 3I)~x = ~u⇔[−4 8−2 4
] [x1x2
]=
[21
]⇔− 2x1 + 4x4 = 1⇔x1 = − 1
2 + 2x2⇔[x1x2
]=
[− 1
2 + 2x2x2
]
⇒ A generalized eigenvector ~v =
[− 1
20
].
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u),
~x(t) = C1e−3t
[21
]+ C2e
−3t([− 1
20
]+ t
[21
])
Example 3 (a)d~x
dt=
[−7 8−2 1
]~x
I Eigenvalues of A: λ1 = λ2 = −3
I An eigenvector for λ1 = λ2 = −3: ~u =
[21
]I Find a generalized eigenvector, by solving (A− λ1I)~x = ~u:
(A+ 3I)~x = ~u⇔[−4 8−2 4
] [x1x2
]=
[21
]⇔− 2x1 + 4x4 = 1⇔x1 = − 1
2 + 2x2⇔[x1x2
]=
[− 1
2 + 2x2x2
]
⇒ A generalized eigenvector ~v =
[− 1
20
].
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u),
~x(t) = C1e−3t
[21
]+ C2e
−3t([− 1
20
]+ t
[21
])
Example 3 (a)d~x
dt=
[−7 8−2 1
]~x
I Eigenvalues of A: λ1 = λ2 = −3
I An eigenvector for λ1 = λ2 = −3: ~u =
[21
]I Find a generalized eigenvector, by solving (A− λ1I)~x = ~u:
(A+ 3I)~x = ~u⇔[−4 8−2 4
] [x1x2
]=
[21
]⇔− 2x1 + 4x4 = 1⇔x1 = − 1
2 + 2x2⇔[x1x2
]=
[− 1
2 + 2x2x2
]⇒ A generalized eigenvector ~v =
[− 1
20
].
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u),
~x(t) = C1e−3t
[21
]+ C2e
−3t([− 1
20
]+ t
[21
])
Example 3 (a)d~x
dt=
[−7 8−2 1
]~x
I Eigenvalues of A: λ1 = λ2 = −3
I An eigenvector for λ1 = λ2 = −3: ~u =
[21
]I Find a generalized eigenvector, by solving (A− λ1I)~x = ~u:
(A+ 3I)~x = ~u⇔[−4 8−2 4
] [x1x2
]=
[21
]⇔− 2x1 + 4x4 = 1⇔x1 = − 1
2 + 2x2⇔[x1x2
]=
[− 1
2 + 2x2x2
]⇒ A generalized eigenvector ~v =
[− 1
20
].
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u),
~x(t) = C1e−3t
[21
]+ C2e
−3t([− 1
20
]+ t
[21
])
Example 3 (a)d~x
dt=
[−7 8−2 1
]~x
I Eigenvalues of A: λ1 = λ2 = −3
I An eigenvector for λ1 = λ2 = −3: ~u =
[21
]I Find a generalized eigenvector, by solving (A− λ1I)~x = ~u:
(A+ 3I)~x = ~u⇔[−4 8−2 4
] [x1x2
]=
[21
]⇔− 2x1 + 4x4 = 1⇔x1 = − 1
2 + 2x2⇔[x1x2
]=
[− 1
2 + 2x2x2
]⇒ A generalized eigenvector ~v =
[− 1
20
].
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u),
~x(t) = C1e−3t
[21
]+ C2e
−3t([− 1
20
]+ t
[21
])
Example 3 (b) Solved~x
dt=
[−7 8−2 1
]~x, ~x(0) =
[23
]
I General solutions:
~x(t) = C1e−3t
[21
]+ C2e
−3t([− 1
20
]+ t
[21
])I Use the initial condition:
~x(0) =
[23
]⇒ C1
[21
]+ C2
[− 1
20
]=
[23
]⇒
[2 − 1
21 0
] [C1
C2
]=
[23
]⇒
[C1
C2
]=
[38
]I The solution to the initial value problem:
~x(t) = 3e−3t[21
]+ 8e−3t
([− 1
20
]+ t
[21
])= e−3t
[2 + 16t3 + 8t
]
Example 3 (b) Solved~x
dt=
[−7 8−2 1
]~x, ~x(0) =
[23
]
I General solutions:
~x(t) = C1e−3t
[21
]+ C2e
−3t([− 1
20
]+ t
[21
])
I Use the initial condition:
~x(0) =
[23
]⇒ C1
[21
]+ C2
[− 1
20
]=
[23
]⇒
[2 − 1
21 0
] [C1
C2
]=
[23
]⇒
[C1
C2
]=
[38
]I The solution to the initial value problem:
~x(t) = 3e−3t[21
]+ 8e−3t
([− 1
20
]+ t
[21
])= e−3t
[2 + 16t3 + 8t
]
Example 3 (b) Solved~x
dt=
[−7 8−2 1
]~x, ~x(0) =
[23
]
I General solutions:
~x(t) = C1e−3t
[21
]+ C2e
−3t([− 1
20
]+ t
[21
])I Use the initial condition:
~x(0) =
[23
]⇒ C1
[21
]+ C2
[− 1
20
]=
[23
]⇒
[2 − 1
21 0
] [C1
C2
]=
[23
]⇒
[C1
C2
]=
[38
]
I The solution to the initial value problem:
~x(t) = 3e−3t[21
]+ 8e−3t
([− 1
20
]+ t
[21
])= e−3t
[2 + 16t3 + 8t
]
Example 3 (b) Solved~x
dt=
[−7 8−2 1
]~x, ~x(0) =
[23
]
I General solutions:
~x(t) = C1e−3t
[21
]+ C2e
−3t([− 1
20
]+ t
[21
])I Use the initial condition:
~x(0) =
[23
]⇒ C1
[21
]+ C2
[− 1
20
]=
[23
]⇒
[2 − 1
21 0
] [C1
C2
]=
[23
]⇒
[C1
C2
]=
[38
]I The solution to the initial value problem:
~x(t) = 3e−3t[21
]+ 8e−3t
([− 1
20
]+ t
[21
])= e−3t
[2 + 16t3 + 8t
]
Example 3 (c) Phase portrait ofd~x
dt=
[−7 8−2 1
]~x
General solutions: ~x(t) = C1e−3t
[21
]+ C2e
−3t([− 1
20
]+ t
[21
])
• When C2 = 0, ~x(t) = C1e−3t
[21
]decays to the origin,
along the eigenspace of λ1 = λ2 = −3.
Example 3 (c) Phase portrait ofd~x
dt=
[−7 8−2 1
]~x
General solutions: ~x(t) = C1e−3t
[21
]+ C2e
−3t([− 1
20
]+ t
[21
])• When C2 = 0, ~x(t) = C1e
−3t[21
]decays to the origin,
along the eigenspace of λ1 = λ2 = −3.
Example 3 (c) Phase portrait ofd~x
dt=
[−7 8−2 1
]~x
General solutions: ~x(t) = C1e−3t
[21
]+ C2e
−3t([− 1
20
]+ t
[21
])• When C2 = 0, ~x(t) = C1e
−3t[21
]decays to the origin,
along the eigenspace of λ1 = λ2 = −3.
Example 3 (c) Phase portrait ofd~x
dt=
[−7 8−2 1
]~x
General solutions: ~x(t) = C1e−3t
[21
]+ C2e
−3t([− 1
20
]+ t
[21
])• When C2 6= 0, ~x(t) ≈ C2te
−3t[21
]for t ≈ ∞,
decaying to the origin along the eigenspace of λ1 = λ2 = −3.
(d) Stability or instability?
The equilibrium (0, 0) isasymptotically stable.
We have anattractive degenerate node,when λ1 = λ2 < 0, but A 6= λ1I.
Example 3 (c) Phase portrait ofd~x
dt=
[−7 8−2 1
]~x
General solutions: ~x(t) = C1e−3t
[21
]+ C2e
−3t([− 1
20
]+ t
[21
])• When C2 6= 0, ~x(t) ≈ C2te
−3t[21
]for t ≈ ∞,
decaying to the origin along the eigenspace of λ1 = λ2 = −3.
(d) Stability or instability?
The equilibrium (0, 0) isasymptotically stable.
We have anattractive degenerate node,when λ1 = λ2 < 0, but A 6= λ1I.
Example 3 (c) Phase portrait ofd~x
dt=
[−7 8−2 1
]~x
General solutions: ~x(t) = C1e−3t
[21
]+ C2e
−3t([− 1
20
]+ t
[21
])• When C2 6= 0, ~x(t) ≈ C2te
−3t[21
]for t ≈ ∞,
decaying to the origin along the eigenspace of λ1 = λ2 = −3.
(d) Stability or instability?
The equilibrium (0, 0) isasymptotically stable.
We have anattractive degenerate node,when λ1 = λ2 < 0, but A 6= λ1I.
Example 3 (c) Phase portrait ofd~x
dt=
[−7 8−2 1
]~x
General solutions: ~x(t) = C1e−3t
[21
]+ C2e
−3t([− 1
20
]+ t
[21
])• When C2 6= 0, ~x(t) ≈ C2te
−3t[21
]for t ≈ ∞,
decaying to the origin along the eigenspace of λ1 = λ2 = −3.
(d) Stability or instability?
The equilibrium (0, 0) isasymptotically stable.
We have anattractive degenerate node,when λ1 = λ2 < 0, but A 6= λ1I.
Example 3 (c) Phase portrait ofd~x
dt=
[−7 8−2 1
]~x
General solutions: ~x(t) = C1e−3t
[21
]+ C2e
−3t([− 1
20
]+ t
[21
])• When C2 6= 0, ~x(t) ≈ C2te
−3t[21
]for t ≈ ∞,
decaying to the origin along the eigenspace of λ1 = λ2 = −3.
(d) Stability or instability?
The equilibrium (0, 0) isasymptotically stable.
We have anattractive degenerate node,when λ1 = λ2 < 0, but A 6= λ1I.
Example 4. (repulsive degenerate node)
Consider ~x′ = A~x, where A =
[1 −14 5
].
(a) Find general solutions of ~x′ =
[1 −14 5
]~x.
(b) Solve the initial value problem ~x′ =
[1 −14 5
]~x, ~x(0) =
[23
].
(c) Sketch the phase portrait.
(d) Is the equilibrium (0, 0) stable, asymptotically stable, or unstable?
Example 4 (a) ~x′ =
[1 −14 5
]~x
I Eigenvalues of A, by solving det(A− λI) = 0:
det
[1− λ −1
4 5− λ
]= λ2 − 6λ+ 9 = 0 ⇒ λ1 = λ2 = 3
I Eigenvectors of A for λ1 = λ2 = 3, by solving (A− λ1I)~x = 0:
(A− 3I)~x = 0⇔[−2 −14 2
] [x1x2
]=
[00
]⇔
[x1x2
]= x2
[− 1
21
]
• Can only pick one linear indep eigenvector ~u =
[− 1
21
].
• Partial solutions: ~x(t) = Ce3t[− 1
21
].
• Need more to get complete solution formula.
Example 4 (a) ~x′ =
[1 −14 5
]~x
I Eigenvalues of A, by solving det(A− λI) = 0:
det
[1− λ −1
4 5− λ
]= λ2 − 6λ+ 9 = 0 ⇒ λ1 = λ2 = 3
I Eigenvectors of A for λ1 = λ2 = 3, by solving (A− λ1I)~x = 0:
(A− 3I)~x = 0⇔[−2 −14 2
] [x1x2
]=
[00
]⇔
[x1x2
]= x2
[− 1
21
]
• Can only pick one linear indep eigenvector ~u =
[− 1
21
].
• Partial solutions: ~x(t) = Ce3t[− 1
21
].
• Need more to get complete solution formula.
Example 4 (a) ~x′ =
[1 −14 5
]~x
I Eigenvalues of A, by solving det(A− λI) = 0:
det
[1− λ −1
4 5− λ
]= λ2 − 6λ+ 9 = 0 ⇒ λ1 = λ2 = 3
I Eigenvectors of A for λ1 = λ2 = 3, by solving (A− λ1I)~x = 0:
(A− 3I)~x = 0⇔[−2 −14 2
] [x1x2
]=
[00
]⇔
[x1x2
]= x2
[− 1
21
]
• Can only pick one linear indep eigenvector ~u =
[− 1
21
].
• Partial solutions: ~x(t) = Ce3t[− 1
21
].
• Need more to get complete solution formula.
Example 4 (a) ~x′ =
[1 −14 5
]~x
I Eigenvalues of A, by solving det(A− λI) = 0:
det
[1− λ −1
4 5− λ
]= λ2 − 6λ+ 9 = 0 ⇒ λ1 = λ2 = 3
I Eigenvectors of A for λ1 = λ2 = 3, by solving (A− λ1I)~x = 0:
(A− 3I)~x = 0⇔[−2 −14 2
] [x1x2
]=
[00
]⇔
[x1x2
]= x2
[− 1
21
]• Can only pick one linear indep eigenvector ~u =
[− 1
21
].
• Partial solutions: ~x(t) = Ce3t[− 1
21
].
• Need more to get complete solution formula.
Example 4 (a)d~x
dt=
[1 −14 5
]~x
I Eigenvalues of A: λ1 = λ2 = 3
I An eigenvector for λ1 = λ2 = 3: ~u =
[− 1
21
]
I Find a generalized eigenvector, by solving (A− λ1I)~x = ~u:
(A− 3I)~x = ~u⇔[−2 −14 2
] [x1x2
]=
[− 1
21
]⇔− 2x1 − x2 = − 1
2⇔x1 = 14 −
12x2⇔
[x1x2
]=
[14 −
12x2
x2
]
⇒ A generalized eigenvector ~v =
[140
].
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u),
~x(t) = C1e3t
[− 1
21
]+ C2e
3t
([140
]+ t
[− 1
21
])
Example 4 (a)d~x
dt=
[1 −14 5
]~x
I Eigenvalues of A: λ1 = λ2 = 3
I An eigenvector for λ1 = λ2 = 3: ~u =
[− 1
21
]I Find a generalized eigenvector, by solving (A− λ1I)~x = ~u:
(A− 3I)~x = ~u⇔[−2 −14 2
] [x1x2
]=
[− 1
21
]⇔− 2x1 − x2 = − 1
2⇔x1 = 14 −
12x2⇔
[x1x2
]=
[14 −
12x2
x2
]
⇒ A generalized eigenvector ~v =
[140
].
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u),
~x(t) = C1e3t
[− 1
21
]+ C2e
3t
([140
]+ t
[− 1
21
])
Example 4 (a)d~x
dt=
[1 −14 5
]~x
I Eigenvalues of A: λ1 = λ2 = 3
I An eigenvector for λ1 = λ2 = 3: ~u =
[− 1
21
]I Find a generalized eigenvector, by solving (A− λ1I)~x = ~u:
(A− 3I)~x = ~u⇔[−2 −14 2
] [x1x2
]=
[− 1
21
]⇔− 2x1 − x2 = − 1
2⇔x1 = 14 −
12x2⇔
[x1x2
]=
[14 −
12x2
x2
]⇒ A generalized eigenvector ~v =
[140
].
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u),
~x(t) = C1e3t
[− 1
21
]+ C2e
3t
([140
]+ t
[− 1
21
])
Example 4 (a)d~x
dt=
[1 −14 5
]~x
I Eigenvalues of A: λ1 = λ2 = 3
I An eigenvector for λ1 = λ2 = 3: ~u =
[− 1
21
]I Find a generalized eigenvector, by solving (A− λ1I)~x = ~u:
(A− 3I)~x = ~u⇔[−2 −14 2
] [x1x2
]=
[− 1
21
]⇔− 2x1 − x2 = − 1
2⇔x1 = 14 −
12x2⇔
[x1x2
]=
[14 −
12x2
x2
]⇒ A generalized eigenvector ~v =
[140
].
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u),
~x(t) = C1e3t
[− 1
21
]+ C2e
3t
([140
]+ t
[− 1
21
])
Example 4 (a)d~x
dt=
[1 −14 5
]~x
I Eigenvalues of A: λ1 = λ2 = 3
I An eigenvector for λ1 = λ2 = 3: ~u =
[− 1
21
]I Find a generalized eigenvector, by solving (A− λ1I)~x = ~u:
(A− 3I)~x = ~u⇔[−2 −14 2
] [x1x2
]=
[− 1
21
]⇔− 2x1 − x2 = − 1
2⇔x1 = 14 −
12x2⇔
[x1x2
]=
[14 −
12x2
x2
]⇒ A generalized eigenvector ~v =
[140
].
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u),
~x(t) = C1e3t
[− 1
21
]+ C2e
3t
([140
]+ t
[− 1
21
])
Example 4 (b) Solved~x
dt=
[1 −14 5
]~x, ~x(0) =
[23
]
I General solutions:
~x(t) = C1e3t
[− 1
21
]+ C2e
3t
([140
]+ t
[− 1
21
])I Use the initial condition:
~x(0) =
[23
]⇒ C1
[− 1
21
]+ C2
[140
]=
[23
]⇒
[− 1
214
1 0
] [C1
C2
]=
[23
]⇒
[C1
C2
]=
[314
]I The solution to the initial value problem:
~x(t) = 3e3t[− 1
21
]+ 14e3t
([140
]+ t
[− 1
21
])= e3t
[2− 7t3 + 14t
]
Example 4 (b) Solved~x
dt=
[1 −14 5
]~x, ~x(0) =
[23
]
I General solutions:
~x(t) = C1e3t
[− 1
21
]+ C2e
3t
([140
]+ t
[− 1
21
])
I Use the initial condition:
~x(0) =
[23
]⇒ C1
[− 1
21
]+ C2
[140
]=
[23
]⇒
[− 1
214
1 0
] [C1
C2
]=
[23
]⇒
[C1
C2
]=
[314
]I The solution to the initial value problem:
~x(t) = 3e3t[− 1
21
]+ 14e3t
([140
]+ t
[− 1
21
])= e3t
[2− 7t3 + 14t
]
Example 4 (b) Solved~x
dt=
[1 −14 5
]~x, ~x(0) =
[23
]
I General solutions:
~x(t) = C1e3t
[− 1
21
]+ C2e
3t
([140
]+ t
[− 1
21
])I Use the initial condition:
~x(0) =
[23
]⇒ C1
[− 1
21
]+ C2
[140
]=
[23
]⇒
[− 1
214
1 0
] [C1
C2
]=
[23
]⇒
[C1
C2
]=
[314
]
I The solution to the initial value problem:
~x(t) = 3e3t[− 1
21
]+ 14e3t
([140
]+ t
[− 1
21
])= e3t
[2− 7t3 + 14t
]
Example 4 (b) Solved~x
dt=
[1 −14 5
]~x, ~x(0) =
[23
]
I General solutions:
~x(t) = C1e3t
[− 1
21
]+ C2e
3t
([140
]+ t
[− 1
21
])I Use the initial condition:
~x(0) =
[23
]⇒ C1
[− 1
21
]+ C2
[140
]=
[23
]⇒
[− 1
214
1 0
] [C1
C2
]=
[23
]⇒
[C1
C2
]=
[314
]I The solution to the initial value problem:
~x(t) = 3e3t[− 1
21
]+ 14e3t
([140
]+ t
[− 1
21
])= e3t
[2− 7t3 + 14t
]
Example 4 (c) Phase portrait ofd~x
dt=
[1 −14 5
]~x
General solutions: ~x(t) = C1e3t
[− 1
21
]+ C2e
3t
([140
]+ t
[− 1
21
])
• When C2 = 0, ~x(t) = C1e3t
[− 1
21
]leaves the origin,
along the eigenspace of λ1 = λ2 = 3.
Example 4 (c) Phase portrait ofd~x
dt=
[1 −14 5
]~x
General solutions: ~x(t) = C1e3t
[− 1
21
]+ C2e
3t
([140
]+ t
[− 1
21
])• When C2 = 0, ~x(t) = C1e
3t
[− 1
21
]leaves the origin,
along the eigenspace of λ1 = λ2 = 3.
Example 4 (c) Phase portrait ofd~x
dt=
[1 −14 5
]~x
General solutions: ~x(t) = C1e3t
[− 1
21
]+ C2e
3t
([140
]+ t
[− 1
21
])• When C2 = 0, ~x(t) = C1e
3t
[− 1
21
]leaves the origin,
along the eigenspace of λ1 = λ2 = 3.
Example 4 (c) Phase portrait ofd~x
dt=
[1 −14 5
]~x
General solutions: ~x(t) = C1e3t
[− 1
21
]+ C2e
3t
([140
]+ t
[− 1
21
])• When C2 6= 0, ~x(t) ≈ C2te
3t
[− 1
21
]is very large, for t ≈ ∞;
~x(t) ≈ C2te3t
[− 1
21
]is very small, for t ≈ −∞.
(d) Stability or instability?
The equilibrium (0, 0) isunstable.
We have a repulsive degenerate node,when λ1 = λ2 > 0, but A 6= λ1I.
Example 4 (c) Phase portrait ofd~x
dt=
[1 −14 5
]~x
General solutions: ~x(t) = C1e3t
[− 1
21
]+ C2e
3t
([140
]+ t
[− 1
21
])• When C2 6= 0, ~x(t) ≈ C2te
3t
[− 1
21
]is very large, for t ≈ ∞;
~x(t) ≈ C2te3t
[− 1
21
]is very small, for t ≈ −∞.
(d) Stability or instability?
The equilibrium (0, 0) isunstable.
We have a repulsive degenerate node,when λ1 = λ2 > 0, but A 6= λ1I.
Example 4 (c) Phase portrait ofd~x
dt=
[1 −14 5
]~x
General solutions: ~x(t) = C1e3t
[− 1
21
]+ C2e
3t
([140
]+ t
[− 1
21
])• When C2 6= 0, ~x(t) ≈ C2te
3t
[− 1
21
]is very large, for t ≈ ∞;
~x(t) ≈ C2te3t
[− 1
21
]is very small, for t ≈ −∞.
(d) Stability or instability?
The equilibrium (0, 0) isunstable.
We have a repulsive degenerate node,when λ1 = λ2 > 0, but A 6= λ1I.
Example 4 (c) Phase portrait ofd~x
dt=
[1 −14 5
]~x
General solutions: ~x(t) = C1e3t
[− 1
21
]+ C2e
3t
([140
]+ t
[− 1
21
])• When C2 6= 0, ~x(t) ≈ C2te
3t
[− 1
21
]is very large, for t ≈ ∞;
~x(t) ≈ C2te3t
[− 1
21
]is very small, for t ≈ −∞.
(d) Stability or instability?
The equilibrium (0, 0) isunstable.
We have a repulsive degenerate node,when λ1 = λ2 > 0, but A 6= λ1I.
Example 4 (c) Phase portrait ofd~x
dt=
[1 −14 5
]~x
General solutions: ~x(t) = C1e3t
[− 1
21
]+ C2e
3t
([140
]+ t
[− 1
21
])• When C2 6= 0, ~x(t) ≈ C2te
3t
[− 1
21
]is very large, for t ≈ ∞;
~x(t) ≈ C2te3t
[− 1
21
]is very small, for t ≈ −∞.
(d) Stability or instability?
The equilibrium (0, 0) isunstable.
We have a repulsive degenerate node,when λ1 = λ2 > 0, but A 6= λ1I.
Example 5. (laminated flow)
Consider ~x′ = A~x, where A =
[−6 4−9 6
].
(a) Find general solutions of ~x′ =
[−6 4−9 6
]~x.
(b) Solve the initial value problem ~x′ =
[−6 4−9 6
]~x, ~x(0) =
[23
].
(c) Sketch the phase portrait.
(d) Is the equilibrium (0, 0) stable, asymptotically stable, or unstable?
Example 5 (a) ~x′ =
[−6 4−9 6
]~x
I Eigenvalues of A, by solving det(A− λI) = 0:
det
[−6− λ 4−9 6− λ
]= λ2 = 0 ⇒ λ1 = λ2 = 0
I Eigenvectors of A for λ1 = λ2 = 0, by solving (A− λ1I)~x = 0:
A~x = 0⇔[−6 4−9 6
] [x1x2
]=
[00
]⇔
[x1x2
]= x2
[231
]
• Can only pick one linear indep eigenvector ~u =
[231
].
• Equilibrium solutions: ~x(t) = C
[231
].
• Need more to get complete solution formula.
Example 5 (a) ~x′ =
[−6 4−9 6
]~x
I Eigenvalues of A, by solving det(A− λI) = 0:
det
[−6− λ 4−9 6− λ
]= λ2 = 0 ⇒ λ1 = λ2 = 0
I Eigenvectors of A for λ1 = λ2 = 0, by solving (A− λ1I)~x = 0:
A~x = 0⇔[−6 4−9 6
] [x1x2
]=
[00
]⇔
[x1x2
]= x2
[231
]
• Can only pick one linear indep eigenvector ~u =
[231
].
• Equilibrium solutions: ~x(t) = C
[231
].
• Need more to get complete solution formula.
Example 5 (a) ~x′ =
[−6 4−9 6
]~x
I Eigenvalues of A, by solving det(A− λI) = 0:
det
[−6− λ 4−9 6− λ
]= λ2 = 0 ⇒ λ1 = λ2 = 0
I Eigenvectors of A for λ1 = λ2 = 0, by solving (A− λ1I)~x = 0:
A~x = 0⇔[−6 4−9 6
] [x1x2
]=
[00
]⇔
[x1x2
]= x2
[231
]
• Can only pick one linear indep eigenvector ~u =
[231
].
• Equilibrium solutions: ~x(t) = C
[231
].
• Need more to get complete solution formula.
Example 5 (a) ~x′ =
[−6 4−9 6
]~x
I Eigenvalues of A, by solving det(A− λI) = 0:
det
[−6− λ 4−9 6− λ
]= λ2 = 0 ⇒ λ1 = λ2 = 0
I Eigenvectors of A for λ1 = λ2 = 0, by solving (A− λ1I)~x = 0:
A~x = 0⇔[−6 4−9 6
] [x1x2
]=
[00
]⇔
[x1x2
]= x2
[231
]• Can only pick one linear indep eigenvector ~u =
[231
].
• Equilibrium solutions: ~x(t) = C
[231
].
• Need more to get complete solution formula.
Example 5 (a)d~x
dt=
[−6 4−9 6
]~x
I Eigenvalues of A: λ1 = λ2 = 0
I An eigenvector for λ1 = λ2 = 0: ~u =
[231
]
I Find a generalized eigenvector, by solving (A− λ1I)~x = ~u:
A~x = ~u⇔[−6 4−9 6
] [x1x2
]=
[231
]⇔− 6x1 + 4x2 = 2
3⇔x1 = − 19 + 2
3x2⇔[x1x2
]=
[− 1
9 + 23x2
x2
]
⇒ A generalized eigenvector ~v =
[− 1
90
].
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u),
~x(t) = C1
[231
]+ C2
([− 1
90
]+ t
[231
])
Example 5 (a)d~x
dt=
[−6 4−9 6
]~x
I Eigenvalues of A: λ1 = λ2 = 0
I An eigenvector for λ1 = λ2 = 0: ~u =
[231
]I Find a generalized eigenvector, by solving (A− λ1I)~x = ~u:
A~x = ~u⇔[−6 4−9 6
] [x1x2
]=
[231
]⇔− 6x1 + 4x2 = 2
3⇔x1 = − 19 + 2
3x2⇔[x1x2
]=
[− 1
9 + 23x2
x2
]
⇒ A generalized eigenvector ~v =
[− 1
90
].
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u),
~x(t) = C1
[231
]+ C2
([− 1
90
]+ t
[231
])
Example 5 (a)d~x
dt=
[−6 4−9 6
]~x
I Eigenvalues of A: λ1 = λ2 = 0
I An eigenvector for λ1 = λ2 = 0: ~u =
[231
]I Find a generalized eigenvector, by solving (A− λ1I)~x = ~u:
A~x = ~u⇔[−6 4−9 6
] [x1x2
]=
[231
]⇔− 6x1 + 4x2 = 2
3⇔x1 = − 19 + 2
3x2⇔[x1x2
]=
[− 1
9 + 23x2
x2
]⇒ A generalized eigenvector ~v =
[− 1
90
].
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u),
~x(t) = C1
[231
]+ C2
([− 1
90
]+ t
[231
])
Example 5 (a)d~x
dt=
[−6 4−9 6
]~x
I Eigenvalues of A: λ1 = λ2 = 0
I An eigenvector for λ1 = λ2 = 0: ~u =
[231
]I Find a generalized eigenvector, by solving (A− λ1I)~x = ~u:
A~x = ~u⇔[−6 4−9 6
] [x1x2
]=
[231
]⇔− 6x1 + 4x2 = 2
3⇔x1 = − 19 + 2
3x2⇔[x1x2
]=
[− 1
9 + 23x2
x2
]⇒ A generalized eigenvector ~v =
[− 1
90
].
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u),
~x(t) = C1
[231
]+ C2
([− 1
90
]+ t
[231
])
Example 5 (a)d~x
dt=
[−6 4−9 6
]~x
I Eigenvalues of A: λ1 = λ2 = 0
I An eigenvector for λ1 = λ2 = 0: ~u =
[231
]I Find a generalized eigenvector, by solving (A− λ1I)~x = ~u:
A~x = ~u⇔[−6 4−9 6
] [x1x2
]=
[231
]⇔− 6x1 + 4x2 = 2
3⇔x1 = − 19 + 2
3x2⇔[x1x2
]=
[− 1
9 + 23x2
x2
]⇒ A generalized eigenvector ~v =
[− 1
90
].
I General solutions are ~x(t) = C1eλ1t~u + C2e
λ1t(~v + t~u),
~x(t) = C1
[231
]+ C2
([− 1
90
]+ t
[231
])
Example 5 (b) Solved~x
dt=
[−6 4−9 6
]~x, ~x(0) =
[23
]
I General solutions:
~x(t) = C1
[231
]+ C2
([− 1
90
]+ t
[231
])I Use the initial condition:
~x(0) =
[23
]⇒ C1
[231
]+ C2
[− 1
90
]=
[23
]⇒
[23 − 1
91 0
] [C1
C2
]=
[23
]⇒
[C1
C2
]=
[30
]I The solution to the initial value problem:
~x(t) = 3
[231
]=
[23
](an equilibrium)
Example 5 (b) Solved~x
dt=
[−6 4−9 6
]~x, ~x(0) =
[23
]
I General solutions:
~x(t) = C1
[231
]+ C2
([− 1
90
]+ t
[231
])
I Use the initial condition:
~x(0) =
[23
]⇒ C1
[231
]+ C2
[− 1
90
]=
[23
]⇒
[23 − 1
91 0
] [C1
C2
]=
[23
]⇒
[C1
C2
]=
[30
]I The solution to the initial value problem:
~x(t) = 3
[231
]=
[23
](an equilibrium)
Example 5 (b) Solved~x
dt=
[−6 4−9 6
]~x, ~x(0) =
[23
]
I General solutions:
~x(t) = C1
[231
]+ C2
([− 1
90
]+ t
[231
])I Use the initial condition:
~x(0) =
[23
]⇒ C1
[231
]+ C2
[− 1
90
]=
[23
]⇒
[23 − 1
91 0
] [C1
C2
]=
[23
]⇒
[C1
C2
]=
[30
]
I The solution to the initial value problem:
~x(t) = 3
[231
]=
[23
](an equilibrium)
Example 5 (b) Solved~x
dt=
[−6 4−9 6
]~x, ~x(0) =
[23
]
I General solutions:
~x(t) = C1
[231
]+ C2
([− 1
90
]+ t
[231
])I Use the initial condition:
~x(0) =
[23
]⇒ C1
[231
]+ C2
[− 1
90
]=
[23
]⇒
[23 − 1
91 0
] [C1
C2
]=
[23
]⇒
[C1
C2
]=
[30
]I The solution to the initial value problem:
~x(t) = 3
[231
]=
[23
](an equilibrium)
Example 5 (c) Phase portrait ofd~x
dt=
[−6 4−9 6
]~x
General solutions:
~x(t) = C1
[231
]+ C2
([− 1
90
]+ t
[231
])
• When C2 = 0,
~x(t) = C1
[231
]are equilibria,
lining along the eigenspace.
Example 5 (c) Phase portrait ofd~x
dt=
[−6 4−9 6
]~x
General solutions:
~x(t) = C1
[231
]+ C2
([− 1
90
]+ t
[231
])
• When C2 = 0,
~x(t) = C1
[231
]are equilibria,
lining along the eigenspace.
Example 5 (c) Phase portrait ofd~x
dt=
[−6 4−9 6
]~x
General solutions:
~x(t) = C1
[231
]+ C2
([− 1
90
]+ t
[231
])
• When C2 = 0,
~x(t) = C1
[231
]are equilibria,
lining along the eigenspace.
Example 5 (c) Phase portrait ofd~x
dt=
[−6 4−9 6
]~x
General solutions:
~x(t) = C1
[231
]+C2
([− 1
90
]+ t
[231
])=
(C1
[231
]+ C2
[− 1
90
])+tC2
[231
]
• When C2 6= 0,
~x(t) are linear functions;
~x(0) = C1
[231
]+ C2
[− 1
90
];
velocity = C2
[231
]parallel to
the eigenspace.
Example 5 (c) Phase portrait ofd~x
dt=
[−6 4−9 6
]~x
General solutions:
~x(t) = C1
[231
]+C2
([− 1
90
]+ t
[231
])=
(C1
[231
]+ C2
[− 1
90
])+tC2
[231
]
• When C2 6= 0,
~x(t) are linear functions;
~x(0) = C1
[231
]+ C2
[− 1
90
];
velocity = C2
[231
]parallel to
the eigenspace.
Example 5 (d) Is the equilibrium (0, 0) stable,asymptotically stable, or unstable?
General solutions: ~x(t) = C1
[231
]+ C2
([− 1
90
]+ t
[231
])
The equilibrium (0, 0) isunstable.
We have a laminar flow,when λ1 = λ2 = 0, but A 6= 0.
Example 5 (d) Is the equilibrium (0, 0) stable,asymptotically stable, or unstable?
General solutions: ~x(t) = C1
[231
]+ C2
([− 1
90
]+ t
[231
])
The equilibrium (0, 0) isunstable.
We have a laminar flow,when λ1 = λ2 = 0, but A 6= 0.
Example 5 (d) Is the equilibrium (0, 0) stable,asymptotically stable, or unstable?
General solutions: ~x(t) = C1
[231
]+ C2
([− 1
90
]+ t
[231
])
The equilibrium (0, 0) isunstable.
We have a laminar flow,when λ1 = λ2 = 0, but A 6= 0.