Contents

Routhhurwitz CriterionEdit

The Routh-Hurwitz criterion is a foundational result in the mathematics of stability, with wide-ranging applications in control theory, signal processing, and the analysis of dynamical systems. It provides a practical test—directly from the coefficients of a real-coefficient polynomial—for whether all roots lie in the left half of the complex plane. In control language, this means whether a system's natural response decays over time rather than grows without bound. The criterion can be formulated in two closely related ways: via the Hurwitz determinants and via the Routh array. Both approaches translate a spectral question about polynomial roots into a condition on the coefficients.

Historically, the problem of locating polynomial roots relative to the imaginary axis was of central importance in 19th-century mathematics and physics, where stability considerations were paramount in mechanical and electrical systems. The results of Adolphe Hurwitz and, independently, E. J. Routh, gave engineers and mathematicians a concrete, checkable procedure. Today, the Routh-Hurwitz criterion remains a standard tool in the analysis and design of continuous-time linear systems, often taught early in control theory curricula and used in automated stability checks for complex systems polynomials, transfer functions, and state-space representations.

Overview

The criterion applies to real-coefficient polynomials of the form p(s) = a0 s^n + a1 s^{n-1} + a2 s^{n-2} + ... + an, with a0 > 0. The goal is to determine whether all roots of p(s) lie to the left of the imaginary axis (real part negative), which corresponds to a stable response in a corresponding continuous-time dynamical system. There are two equivalent formulations commonly used in practice:

  • Hurwitz determinants: Construct the Hurwitz matrix from the coefficients and form its leading principal determinants Δk for k = 1, 2, ..., n. The system is stable if and only if each Δk > 0.
  • Routh array: Build a Routh table (array) from the coefficients. The system is stable if and only if every element in the first column of the table is positive (assuming a0 > 0). The sign pattern of the first column encodes the location of the roots.

In both formulations, the core idea is to convert a spectral condition on roots into a sign condition on quantities derived from the coefficients. The Routh array is particularly popular for hand calculations and for illuminating the structure of the stability conditions, while the Hurwitz determinants provide a compact, determinant-based view.

Formal statements

Hurwitz determinants

Given p(s) = a0 s^n + a1 s^{n-1} + ... + an with a0 > 0, form the Hurwitz matrix H(p) from the coefficients. The determinants of the top-left k-by-k submatrices, Δk, are called the Hurwitz determinants. The Routh-Hurwitz stability test states that p(s) has all roots with negative real parts if and only if

  • Δk > 0 for all k = 1, 2, ..., n.

This criterion is equivalent to the positivity of the leading principal minors of the Hurwitz matrix. See also Hurwitz matrix for the matrix construction and related linear-algebraic interpretations.

Routh array

Construct the Routh table from the coefficients by arranging them in the first two rows as follows:

  • Row s^n: a0, a2, a4, ...
  • Row s^{n-1}: a1, a3, a5, ...

Subsequent rows are filled using a standard recurrence that depends only on the two rows above. The general rule can be stated as R(i, j) = (R(i-1, 0) * R(i-2, j+1) - R(i-2, 0) * R(i-1, j+1)) / R(i-1, 0),

with the understanding that missing entries are treated as zero. The crucial feature is that the sign of the first column determines stability: all first-column entries must be positive (assuming a0 > 0) for p(s) to be strictly stable. If any first-column entry is zero or negative, there is at least one root with nonnegative real part, indicating instability or marginal stability.

Worked example

Consider the cubic p(s) = s^3 + 2 s^2 + 3 s + 4.

  • Coefficients: a0 = 1, a1 = 2, a2 = 3, a3 = 4.
  • Routh table construction:
    • s^3 row: 1, 3
    • s^2 row: 2, 4
    • s^1 row: (2*3 - 1*4)/2 = 1
    • s^0 row: 4

The first column is [1, 2, 1, 4], all positive, so p(s) is stable (all roots have negative real parts).

A non-stable example is p(s) = s^3 - 2 s^2 + 3 s - 4, with coefficients a0 = 1, a1 = -2, a2 = 3, a3 = -4. The first column begins with [1, -2, ...], which includes a negative entry, signaling at least one root with nonnegative real part and thus instability.

Applications and connections

  • Control theory: Used to assess the stability of closed-loop systems described by continuous-time models, including those expressed as transfer functions transfer functions and, more generally, in state-space representations. The criterion directly links polynomial coefficients to stability without requiring explicit root-finding.
  • System design and verification: Provides a quick check during the design of controllers and observers to ensure that chosen gains and dynamics yield a stable response.
  • Related criteria: The Routh-Hurwitz criterion sits alongside other classical stability tools such as the Nyquist criterion and the Lyapunov stability framework. Each approach has its domain of convenience; the Routh-Hurwitz method excels for polynomial-characterized dynamics, while Nyquist analysis is often preferred when frequency-domain information is readily available.

Extensions and related results

  • Higher-order polynomials and numerical issues: For very high-order systems, the construction of the Routh table or the evaluation of Hurwitz determinants can be numerically sensitive. Modern tools often combine these classical tests with robust numerical methods and marginal-stability analysis.
  • Robustness and parametric uncertainty: In real-world applications, stability must be maintained under parameter variations. Extensions of the Routh-Hurwitz framework cope with uncertainty by combining the criterion with interval arithmetic or certified robustness analyses.
  • Generalizations: The ideas behind the Hurwitz determinants and the Routh array extend to related stability questions in more advanced settings, including discrete-time systems (where the stability region is inside the unit circle rather than the left half-plane) and systems with special structures.

See also