The spectral radius of an operator or matrix is the supremum of the absolute values (moduli) of elements in its spectrum; symbolically r(A)=max{|λ|: λ∈σ(A)} when the maximum exists.

Principle

Spectral radius measures asymptotic growth rates under iteration and is linked to norms via inequalities (r(A) ≤ ||A||) and, for elements of Banach algebras, to Gelfand's formula r(A)=lim_{n→∞}||A^n||^{1/n} when the limit exists.

Demonstration

For a diagonal matrix with diagonal entries 2,−3,1 the spectral radius is 3. For a nilpotent matrix all eigenvalues are 0 and the spectral radius is 0 despite possibly large operator norms for small powers.

Misapplication

Equating operator norm with spectral radius is incorrect: the spectral radius can be strictly smaller than any given operator norm. Using norm bounds as if they were exact spectral values misleads asymptotic predictions.

Consequence

If the spectral radius of A is <1 then I−A is invertible and Neumann series converge; the spectral radius controls long-term behavior of powers A^n and stability of linear dynamical systems and numerical schemes.

Reversal

Considering the inverse notion — the minimal modulus of spectral points — highlights spectral gaps and stability margins; an operator with spectral radius 0 is quasinilpotent (all spectral points at 0) whereas large spectral radius indicates possible growth.

Boundary

Spectral radius is meaningful only once the spectrum is defined and typically considered over C; for infinite-dimensional operators subtleties arise (e.g., approximate point spectrum influencing growth), and equality in Gelfand's formula may require algebraic conditions.

Semantic Tension

Tension exists between spectral radius, numerical radius, and operator norm: all measure size but with different meanings — spectral radius is spectral (asymptotic eigenvalue size), numerical radius relates to quadratic forms, and norm measures operator action on vectors.

Synthesis

The spectral radius is the principal spectral magnitude: the largest modulus of spectral values, governing asymptotic powers, invertibility criteria like Neumann series convergence, and indicating growth or decay tendencies of the operator.