generalized vector v2 = (0, 1)T : • eigenvalue classification

    • node
x 1 x x • saddle
Y = c1 e + c2 e .
0 1 • spiral
• center
Related / Associated Concepts • stable / asymptotically stable / unstable
Associated concepts:
• matrix exponential Definitions
• Jordan form • Autonomous planar system:
• phase plane
ẋ = P (x, y), ẏ = Q(x, y).
• autonomous systems
• stability • Critical point / equilibrium point (x0 , y0 ) satisfies
Core Concept Skeleton for Related Concepts P (x0 , y0 ) = 0, Q(x0 , y0 ) = 0.
• eigenspace
• generalized eigenchain • Stable: trajectories starting sufficiently close remain
• trace close.
• determinant • Asymptotically stable: stable and trajectories suffi-
ciently close converge to equilibrium.
• invariant subspace
• Unstable: not stable.
Direct Results / Formulae for Related Con- • Jacobian at (x0 , y0 ):
cepts  
Px P y
J= .
• Characteristic polynomial in 2 × 2 case: Qx Qy (x ,y )
0 0

2
λ − (trA)λ + det A = 0.
Theorems / Direct Statements
• For diagonalizable A = P DP −1
: • Linearized system near equilibrium:
eAx = P eDx P −1 . U̇ = JU.
• If A = λI + N with N k = 0: • For linear system U̇ = AU with characteristic equation
x2 N 2
 
eAx = eλx I + xN + + ··· . λ2 − (trA)λ + det A = 0,
2!
classification is:
– real distinct same sign ⇒ node;
CSIR NET Utility Layer
– real distinct opposite sign ⇒ saddle;
• In CSIR problems, constant-coefficient systems are – real repeated ⇒ node / improper node;
usually solved fastest by eigenvalues. – complex with nonzero real part ⇒ spiral;
• If matrix is triangular, eigenvalues are diagonal entries. – purely imaginary ⇒ center for linear system.
• A repeated eigenvalue needs a quick check: one eigen- • Stability for linear planar system:
vector or two? – both eigenvalues with negative real parts ⇒ asymp-
• Variation of parameters in matrix form is high-yield totically stable;
but usually statement-based. – at least one eigenvalue with positive real part ⇒
unstable;
Research-Based Enrichment – purely imaginary eigenvalues ⇒ stable but not
• Fundamental matrices satisfy the cocycle relation asymptotically stable in the linear case.
Φ(x, s)Φ(s, t) = Φ(x, t). • For nonlinear planar systems, if eigenvalues of Jaco-
• The determinant evolution is governed only by trA, bian have nonzero real parts, local type matches the
not by the full matrix. linearization (hyperbolic case).
• Systems theory links directly with semigroup methods
and linear algebraic canonical forms. Direct Results / Formulae
• Trace-determinant data:
9. Stability Analysis and Critical Points for
Planar Systems τ = trA, ∆ = det A, D = τ 2 − 4∆.

Core Concept Skeleton • Quick classification:

• autonomous system – ∆ < 0: saddle;
• equilibrium point – ∆ > 0, D > 0, τ ̸= 0: node;
• linearization – ∆ > 0, D < 0, τ ̸= 0: spiral/focus;
• Jacobian – ∆ > 0, D < 0, τ = 0: center (linear);

12
 – ∆ > 0, D = 0: repeated-root node/star/improper Direct Results / Formulae for Related Con-
node. cepts
• Stability:
– τ < 0, ∆ > 0 and not center-type ⇒ asymptotically • Nonlinear system near equilibrium (x0 , y0 ):
stable;
U̇ = JU + higher order terms.
– τ > 0, ∆ > 0 ⇒ unstable;
– ∆ < 0 ⇒ unstable. • Hyperbolic condition:

ℜ(λi ) ̸= 0 ∀i.

Alternative / Equivalent / Easier Forms Then topological type is preserved under nonlinear
perturbation near equilibrium.
• Node = real same-sign eigenvalues.
• Saddle = determinant negative.
• Center = trace zero, determinant positive, discriminant CSIR NET Utility Layer
negative for the linear system. • Fastest rule: compute determinant first. If negative, it
• Spiral = complex eigenvalues with nonzero real part. is a saddle and hence unstable.
• In trace-determinant plane, parabola τ 2 = 4∆ sepa- • Trace decides stable vs unstable when determinant is
rates real and complex eigenvalues. positive.
• In objective questions, purely imaginary immediately
suggests center for the linear system.
• “Same type as linearization” is valid only for hyperbolic
Failure Cases / Exceptions / Warnings critical points.
• Purely imaginary eigenvalues in nonlinear systems are
inconclusive; linearization may fail. Research-Based Enrichment
• Repeated real eigenvalues do not alone determine ge- • Planar linear stability is the finite-dimensional ana-
ometry; need eigenspace dimension. logue of spectral stability in PDEs.
• Center in linearization can become center or spiral in • Lyapunov functions provide nonlinear stability tools
nonlinear system. beyond Jacobian criteria.
• Hyperbolicity fails when an eigenvalue has zero real • Center manifolds explain why purely imaginary / zero-
part. real-part cases are delicate.

10. Boundary Value Problems

Examples Core Concept Skeleton
• Basic: • IVP vs BVP
ẋ = −x, ẏ = −2y. • second-order linear operator
Eigenvalues −1, −2; asymptotically stable node. • homogeneous boundary conditions
• Tricky: • uniqueness criterion
ẋ = y, ẏ = −x. • associated homogeneous problem
  • self-adjoint form
0 1
Matrix has eigenvalues ±i; linear center, • eigenvalue problem
−1 0
• Sturm–Liouville structure
stable but not asymptotically stable.
• orthogonality
• regular / singular / periodic

Related / Associated Concepts Definitions

Associated concepts: • Standard second-order BVP:
• phase portrait y ′′ + P (x)y ′ + Q(x)y = ϕ(x), a < x < b,
• Jacobian linearization
with boundary conditions
• hyperbolic equilibrium
• Lyapunov stability α1 y(a) + β1 y ′ (a) = γ1 , α2 y(b) + β2 y ′ (b) = γ2 .
• trace-determinant plane
• Homogeneous BVP: ϕ ≡ 0 and γ1 = γ2 = 0.
Core Concept Skeleton for Related Concepts • Eigenvalue problem:
• local flow d
(p(x)y ′ ) + [q(x) + λr(x)]y = 0
• eigen-directions dx
• invariant manifolds with homogeneous boundary conditions.
• sign of real part • Regular Sturm–Liouville problem:
• nonlinear perturbation
−(py ′ )′ + qy = λry, p > 0, r > 0, x ∈ [a, b],

13
 with separated homogeneous boundary conditions. checking invertibility of the boundary-condition matrix
after writing the general solution.
Theorems / Direct Statements • Regular, singular, periodic are classifications by end-
points/boundary conditions.
• A non-homogeneous linear BVP has a unique solution
iff the associated homogeneous BVP has only the trivial
solution.
• If y1 , y2 are linearly independent solutions of the homo-
geneous differential equation, then nontrivial solutions
of the homogeneous BVP exist iff the boundary deter-
minant vanishes. Failure Cases / Exceptions / Warnings
• For regular/self-adjoint Sturm–Liouville problems: • Unlike IVPs, BVPs may have no solution, one solution,
– all eigenvalues are real; or infinitely many solutions.
– eigenfunctions for distinct eigenvalues are linearly • If the associated homogeneous BVP has nontrivial
independent; solutions, uniqueness fails.
– eigenfunctions for distinct eigenvalues are orthogonal • Self-adjoint spectral conclusions require the correct hy-
with respect to weight r(x): potheses: especially p > 0, r > 0, coefficient continuity,
and admissible boundary conditions.
b
• Periodic BVP eigenvalue structure differs from sepa-
Z
r(x)ym (x)yn (x) dx = 0, m ̸= n.
a rated regular BVP structure.
• For regular problems, eigenvalues form a countably
infinite sequence

λ1 < λ2 < · · · , λn → +∞.

• For regular Sturm–Liouville problems, the nth eigen- Examples

function has exactly (n − 1) zeros in (a, b). • Basic BVP:
• Dimension of each eigenspace for regular separated
Sturm–Liouville problem is 1. y ′′ = 0, y(0) = 0, y(1) = 1.

Direct Results / Formulae General solution y = c1 x + c2 ; boundary conditions

give unique solution y = x.
• Boundary determinant for homogeneous BVP with • Exam-style eigenvalue problem:
homogeneous differential equation and fundamental
solutions y1 , y2 : y ′′ + λy = 0, y(0) = 0, y(π) = 0.
α1 y1 (a) + β1 y1′ (a) α1 y2 (a) + β1 y2′ (a) Nontrivial solutions occur for
∆= .
α2 y1 (b) + β2 y1′ (b) α2 y2 (b) + β2 y2′ (b)
λ = n2 , n ∈ N,
Nontrivial solutions exist iff ∆ = 0.
• Self-adjoint conversion of with eigenfunctions

a0 (x)y ′′ + a1 (x)y ′ + a2 (x)y = 0 yn (x) = sin(nx).

requires
a′0 (x) = a1 (x)
or, after multiplying by a suitable integrating factor,
can be written as
Related / Associated Concepts
d
p(x)y ′ + q(x)y = 0.


dx Associated concepts:
• Sturm–Liouville theory
• Typical periodic boundary conditions:
• orthogonality
y(a) = y(b), y ′ (a) = y ′ (b). • least eigenvalue
• nodal theorem
Alternative / Equivalent / Easier Forms • self-adjoint operators

• The self-adjoint form may be written as either Core Concept Skeleton for Related Concepts
• weighted inner product
(py ′ )′ + (λr + q)y = 0
• endpoint conditions
or • spectrum
−(py ′ )′ + qy = λry. • nodal count
• variational ordering
• Non-homogeneous BVP solvability is often reduced to

14