Stability in Linear Systems - Linear Algebra

  1. xk+1=Axk\mathbf{x}_{k+1} = A\mathbf{x}_k where AA has complex eigenvalues of modulus 0.80.8. What do solutions do?
    They spiral in toward 0\mathbf{0}, since ∣λ∣=0.8<1|\lambda| = 0.8 < 1.
  2. For xk+1=Axk\mathbf{x}_{k+1} = A\mathbf{x}_k, when do all solutions approach 0\mathbf{0}?
    When every eigenvalue of AA satisfies ∣λ∣<1|\lambda| < 1.
  3. xk+1=Axk\mathbf{x}_{k+1} = A\mathbf{x}_k, AA diagonalizable with eigenvalues 11 and 0.40.4. What happens to a typical solution?
    It converges to a fixed multiple of the eigenvector for 11, since the 0.4k0.4^k component dies out.
  4. x′=Ax\mathbf{x}' = A\mathbf{x} with eigenvalues ±2i\pm 2i. What do nonzero solutions do?
    They move periodically on closed orbits around the origin (a center), neither growing nor decaying.
  5. x′=Ax\mathbf{x}' = A\mathbf{x} with eigenvalues 22 and −1-1. What kind of equilibrium is the origin?
    A saddle point, which is unstable: solutions decay along one eigenvector and grow along the other.
  6. x′=Ax\mathbf{x}' = A\mathbf{x} with eigenvalues −1-1 and −3-3. What is the long-term behavior?
    Every solution decays to 0\mathbf{0}; the origin is a stable node (an attractor).
  7. In a discrete system xk+1=Axk\mathbf{x}_{k+1} = A\mathbf{x}_k, what role does the dominant eigenvalue play?
    The eigenvalue of largest absolute value sets the long-run growth rate, and solutions align with its eigenvector.
  8. For a 2×22 \times 2 system x′=Ax\mathbf{x}' = A\mathbf{x} with tr A<0\text{tr}\,A < 0 and det⁡A>0\det A > 0, what is the long-term behavior?
    All solutions approach 0\mathbf{0}: these conditions mean both eigenvalues have negative real parts.
  9. xk+1=Axk\mathbf{x}_{k+1} = A\mathbf{x}_k with eigenvalues 0.90.9 and 0.60.6. What happens to every solution as k→∞k \to \infty?
    It approaches 0\mathbf{0}, since both eigenvalues have absolute value less than 11.
  10. xk+1=Axk\mathbf{x}_{k+1} = A\mathbf{x}_k, AA diagonalizable with eigenvalues 1.21.2 and 0.50.5. What does a typical solution do as k→∞k \to \infty?
    It grows without bound along the eigenvector for 1.21.2. Only starts on the eigenline for 0.50.5 decay to 0\mathbf{0}.
  11. x′=Ax\mathbf{x}' = A\mathbf{x} with eigenvalues 1±2i1 \pm 2i. What do nonzero solutions do?
    They spiral outward away from the origin, since the real part 11 is positive.
  12. For x′=Ax\mathbf{x}' = A\mathbf{x}, when do all solutions approach 0\mathbf{0}?
    When every eigenvalue of AA has negative real part.

Study these flashcards