Introduction to the Routh‑Hurwitz Stability Criterion

The Routh–Hurwitz stability criterion is a cornerstone of control theory and polynomial analysis. It provides a systematic method for determining the number of roots of a real polynomial that lie in the right half of the complex plane—without requiring explicit root computation. For engineers and mathematicians working with complex polynomials, this criterion transforms an otherwise difficult algebraic problem into a straightforward tabular procedure.

A “complex polynomial” here refers to a polynomial with real coefficients (as is standard in control applications) but whose roots may be complex. The stability of many physical systems—from electronic circuits to mechanical linkages—depends on the sign of the real parts of the characteristic polynomial’s roots. The Routh–Hurwitz criterion answers this question efficiently, even for high‑order polynomials that are impractical to factor.

This article expands on the original summary, providing a detailed walkthrough of the method, examples, special cases, and practical applications. By the end, readers will be able to apply the Routh–Hurwitz technique to simplify and analyze complex polynomials with confidence.

What Are Complex Polynomials?

In control engineering and mathematics, a complex polynomial is typically an expression of the form:

P(s) = an sn + an−1 sn−1 + … + a1 s + a0

where the coefficients ai are real numbers and s is a complex variable. The term “complex” does not refer to the coefficients (which remain real) but to the fact that the roots of the equation P(s)=0 are generally complex numbers. These roots determine the dynamic behavior of a system described by the polynomial.

For example, the characteristic equation of a second‑order system like a mass‑spring‑damper is ms2 + bs + k = 0. Depending on the damping ratio, the roots may be real and negative (overdamped) or a complex conjugate pair with negative real part (underdamped). The Routh–Hurwitz criterion provides a way to check whether all roots have negative real parts—without solving the equation.

Why Direct Root Calculation Is Impractical

For polynomials of order five or higher, exact analytical root solutions rarely exist (Abel–Ruffini theorem). Numerical methods can approximate roots, but they are sensitive to coefficient uncertainty and may miss subtle signs of instability. The Routh–Hurwitz criterion, by contrast, gives a definitive yes/no answer about the sign of real parts, using only the polynomial coefficients. This makes it indispensable for preliminary stability analysis and for verifying design margins.

The Routh–Hurwitz Stability Criterion

Developed independently by Edward John Routh (1876) and Adolf Hurwitz (1895), the criterion states that for a polynomial with real coefficients, all roots have negative real parts if and only if all the leading principal minors of the Hurwitz matrix (or, equivalently, the first column of the Routh array) are positive. The method is usually taught using the Routh array because it is simpler to construct by hand.

Constructing the Routh Array

Given the polynomial:

P(s) = an sn + an−1 sn−1 + … + a1 s + a0 (with an > 0),

the Routh array is built as follows:

  1. Label the first two rows. Row sn contains coefficients of even‑indexed powers (starting with an, an−2, …). Row sn−1 contains coefficients of odd‑indexed powers (an−1, an−3, …). Any missing coefficients are filled with zeros.
  2. Compute subsequent rows. For row sn−2, each element is derived from the two rows above using a determinant formula:
    b1 = (an−1 an−2 - an an−3) / an−1,
    b2 = (an−1 an−4 - an an−5) / an−1,
    and so on, until the remaining elements become zero.
  3. Continue until only the constant row remains. The last row corresponds to s0 and contains a single element, which is the constant term (typically a0).
  4. Examine the first column. If all entries in the first column of the array are positive (and none are zero), then all roots lie in the left half‑plane (the system is stable). Any sign change indicates the number of roots with positive real parts—each sign change counts one such root.

Special Cases in the Routh Array

Two special situations arise that require careful handling:

  • Zero in the first column. If the first element of a row is zero, but the rest of the row is nonzero, replace the zero with a small positive number ε and proceed. The sign of ε determines the stability outcome. Alternatively, one can use the auxiliary polynomial method (see below).
  • Entire row of zeros. This indicates that the polynomial contains symmetrically located roots (pairs of roots that are negatives of each other or complex conjugates with zero real part). In this case, form an auxiliary polynomial from the row above the zeros, take its derivative, and use the coefficients of that derivative to replace the zero row. Then continue the array.

Example: A Fourth‑Order Polynomial

Consider P(s) = s4 + 2s3 + 3s2 + 4s + 5.

  • Row s4: 1, 3, 5
  • Row s3: 2, 4, 0
  • Row s2: b1 = (2·3 - 1·4)/2 = (6-4)/2 = 1, b2 = (2·5 - 1·0)/2 = 10/2 = 5
  • Row s1: c1 = (1·4 - 2·5)/1 = (4-10)/1 = -6, then zero
  • Row s0: d1 = (-6·5 - 1·0)/(-6) = (-30)/(-6) = 5

First column: 1, 2, 1, -6, 5 → the sign changes once (from 1 to -6, and from -6 back to 5), indicating one root with positive real part. Indeed, the polynomial has roots: -1.65 ± 0.35i, 0.65 ± 1.72i — the last pair has positive real part.

Simplifying Complex Polynomials Using the Criterion

The Routh–Hurwitz method simplifies the stability analysis of complex polynomials in several ways:

  1. Avoids root computation. For high‑order polynomials, factoring or numerically searching for roots is expensive and sometimes unreliable. The array requires only addition, subtraction, multiplication, and division—operations easily performed by hand or spreadsheet.
  2. Handles parameter uncertainty. By expressing coefficients in terms of system parameters (gains, time constants), one can derive stability conditions in terms of inequalities. For example, the range of a controller gain K for which the closed‑loop system remains stable can be found by applying the Routh array and requiring all first‑column elements to be positive.
  3. Identifies marginal stability. A zero in the first column (the ε case) or a row of zeros indicates roots on the imaginary axis or symmetrically placed about the origin. These conditions often define the boundary between stable and unstable behavior.

Working with Complex Coefficients?

While the classic Routh–Hurwitz criterion requires real coefficients, many extensions exist. For polynomials with complex coefficients (ai ∈ ℂ), modifications such as the modified Routh array or the Michailov criterion can be employed. However, in most practical control engineering problems, the characteristic equation arises from real‑valued physical parameters, so real coefficients are the norm.

Applications in Control Systems and Beyond

The primary application of the Routh–Hurwitz stability criterion is in determining the stability of linear time‑invariant (LTI) systems described by transfer functions. The denominator of the transfer function is the characteristic polynomial, and stability requires all roots to have negative real parts.

Design of Control Systems

During controller design, engineers often adjust gains (e.g., a PID controller’s proportional, integral, and derivative gains) to achieve desired performance. The Routh criterion provides a quick way to find gain margins—the range of gains that keep the system stable. For example, given a unity feedback system with forward transfer function G(s) = K / [(s+1)(s+2)(s+3)], the characteristic polynomial is s3 + 6s2 + 11s + (6+K) = 0. Applying the Routh array yields the condition K < 60 for stability. This is far faster than trying to compute roots as a function of K.

Network Analysis

In electrical engineering, the characteristic polynomial of an RLC circuit determines whether transient oscillations decay. The Routh–Hurwitz criterion helps engineers choose component values to avoid sustained oscillations (instability). Similarly, in mechanical systems, the stability of feedback‑controlled robots or aircraft is routinely assessed using this method.

Industrial Process Control

In chemical and petroleum industries, process controllers regulate temperature, pressure, and flow. The Routh array is used during the tuning of proportional–integral–derivative (PID) controllers to ensure that the closed‑loop plant remains stable under variations in operating conditions.

Benefits and Limitations

Like any tool, the Routh–Hurwitz criterion has strengths and weaknesses.

Strengths

  • No root calculation required. The array operates solely on coefficients.
  • Scalable. The method works for polynomials of any order, though higher orders become tedious by hand (easily automated).
  • Symbolic capability. Coefficients can be algebraic expressions, allowing parametric stability analysis.
  • Provides the number of unstable roots. Each sign change in the first column corresponds to one root with positive real part.

Limitations

  • Real coefficients only. The classic form does not directly handle polynomials with complex coefficients, though extensions exist.
  • No information about root locations. The criterion only tells how many unstable roots exist, not their exact values. For detailed transient analysis, root locus or numerical root‑finding is needed.
  • Sensitivity to numerical errors. When coefficients are nearly zero or contain rounding errors, the array construction can produce sign ambiguities. Use of the ε‑method is a workaround but requires careful interpretation.
  • Does not handle delay systems. Systems with time delays have transcendental characteristic equations, not polynomials. The Routh–Hurwitz criterion is not directly applicable; instead, methods like Padé approximations or the Smith predictor are used.

Historical Context and Further Reading

The Routh–Hurwitz criterion is named after Edward John Routh, an English mathematician who published the array method in 1876, and Adolf Hurwitz, a German mathematician who independently derived a similar matrix formulation in 1895. Their work built on earlier contributions by Cauchy and Hermite. The criterion remains a staple in modern control textbooks—see the Wikipedia article for an overview. For a deeper mathematical treatment, ScienceDirect’s topic page provides references to original sources.

Engineers studying advanced control will also encounter the Hurwitz matrix in the context of Lyapunov stability. The connection between the Routh array and the determinant of the Hurwitz matrix is a classic result in linear algebra and control theory.

Step‑by‑Step Illustrated Example

Let’s walk through a complete example with a complex (high‑order) polynomial to demonstrate the method’s power. Consider the characteristic polynomial of a control system:

P(s) = 2 s5 + 4 s4 + 6 s3 + 8 s2 + 10 s + 12

Step 1: Arrange the first two rows

  • Row s5 (power 5): coefficients of s5, s3, s1 → 2, 6, 10
  • Row s4 (power 4): coefficients of s4, s2, s0 → 4, 8, 12

Step 2: Compute the third row (s3)

Using the determinant formula:

  • First element: (4·6 - 2·8)/4 = (24 - 16)/4 = 8/4 = 2
  • Second element: (4·10 - 2·12)/4 = (40 - 24)/4 = 16/4 = 4
  • Third element: zero (since no more coefficients in the first row)

Row s3: 2, 4, 0

Step 3: Compute the fourth row (s2)

Using row s4 (4,8,12) and row s3 (2,4,0):

  • First element: (2·8 - 4·4)/2 = (16 - 16)/2 = 0/2 = 0
  • Second element: (2·12 - 4·0)/2 = (24 - 0)/2 = 12

Row s2: 0, 12

Here we have a zero in the first column. The rest of the row is nonzero (12). Replace the zero with ε (a small positive number).

Step 4: Compute the fifth row (s1)

Using row s3 (2,4,0) and row s2 (ε,12):

  • First element: (ε·4 - 2·12)/ε = (4ε - 24)/ε = 4 - 24/ε
  • Second element: (ε·0 - 2·0)/ε = 0

Row s1: (4 - 24/ε), 0

As ε → 0+, the term -24/ε dominates, making this element negative (−∞).

Step 5: Compute the sixth row (s0)

Using row s2 (ε,12) and row s1 (4 - 24/ε, 0):

  • First element: [(4 - 24/ε)·12 - ε·0] / (4 - 24/ε) = 12

Row s0: 12

Step 6: Analyze the first column

The first column entries are: row s5: 2 (positive), row s4: 4 (positive), row s3: 2 (positive), row s2: ε (positive, as ε>0), row s1: 4 - 24/ε (negative for small ε), row s0: 12 (positive).

There are two sign changes: from positive (row s2) to negative (row s1), and from negative back to positive (row s0). Therefore, the polynomial has two roots with positive real parts.

Verification: Using numerical tools, the roots of 2 s5 + … + 12 = 0 are approximately: -1.09, -0.50 ± 0.87i, 0.79 ± 1.04i. Indeed, the two complex conjugate roots have positive real part (0.79), confirming the Routh‑Hurwitz prediction.

Alternative approach: The auxiliary polynomial for the zero‑row case

If we had encountered a full row of zeros instead of just a zero in the first column, we would form an auxiliary polynomial from the row above. That technique is essential for cases like s4 + 5s2 + 4 (which has purely imaginary roots ±j, ±2j). The zero row indicates roots symmetric about the origin, and the auxiliary polynomial (derived from the row above) gives the actual roots.

Conclusion

The Routh–Hurwitz stability criterion transforms the complex analysis of high‑order polynomials into a straightforward algorithm. By constructing the Routh array and inspecting the signs of the first column, engineers and mathematicians can determine system stability without solving for roots. The method is indispensable in control system design, circuit analysis, and any field where the roots of a real polynomial dictate behavior.

Mastering this technique allows one to handle complex polynomials with confidence—whether for academic study, industrial design, or research. Combined with modern computational tools (e.g., MATLAB, Python), the Routh‑Hurwitz criterion remains a fundamental building block of dynamic system analysis. For those wishing to explore further, the Control Systems Principles whitepaper provides additional worked examples and advanced extensions.