First Year Engineering is coming soon

We're still writing and reviewing the notes for every subject, and we'll open it when every subject is ready. Meanwhile you can practise aptitude, ask Aria about any subject, or see the branches that are ready.

Euler's theorem on homogeneous functions

Euler's theorem on homogeneous functions: definition of homogeneous functions and degree, Euler’s first- and second-order identities, proofs, extensions to many variables, and exam-standard applications and checks.

Drafted with Aria, reviewed by the AiCanCode.org team · Updated . How we write and check our notes. Spotted an error? Use Give Feedback at the bottom of the page.

Why it matters

Engineers often work with models that scale predictably, such as stresses with size, power laws in materials, and similarity in fluid mechanics. Homogeneous functions capture this scaling: if you scale inputs by a factor, the output scales by a fixed power. Euler’s theorem gives a direct relation between a homogeneous function and its partial derivatives, which simplifies analysis, derivations, and checking results.

Key ideas

  • Examples: x² + y² (degree 2), x³y (degree 4), (x² + y²)^(3/2) (degree 3), any function of ratios alone like f(y/x) is degree 0.

  • Degree of homogeneity (definition): The number n in the definition above is called the degree (or order) of the homogeneous function.

  • Euler’s theorem on homogeneous functions (statement, two variables): Let z = f(x, y) be continuously differentiable and homogeneous of degree n. Then x (∂z/∂x) + y (∂z/∂y) = n z. Assumptions: f has continuous first partial derivatives in a domain excluding the origin if needed; f is homogeneous of degree n in that domain. Converse (two variables): If f is C¹ on a star-shaped domain (excluding the origin if needed) and satisfies x f₍ₓ₎ + y f₍y₎ = n f for all (x, y) in the domain, then f is homogeneous of degree n.

  • Euler’s theorem (many variables): If u = f(x₁, x₂, …, xₖ) is C¹ (continuously differentiable) and homogeneous of degree n, then Σᵢ xᵢ (∂u/∂xᵢ) = n u, where the sum is from i = 1 to k. Moreover, each first partial ∂u/∂xᵢ is homogeneous of degree n − 1.

  • Second-order Euler-type relation (two variables): If z = f(x, y) is C² (has continuous second partials) and homogeneous of degree n, then x² z₍ₓₓ₎ + 2xy z₍ₓy₎ + y² z₍yy₎ = n(n − 1) z. This follows by differentiating Euler’s first-order identity.

  • Special case: degree zero: If u = f(x, y) is homogeneous of degree 0 (e.g., f depends only on y/x), then x u₍ₓ₎ + y u₍y₎ = 0. This is often used to show invariance under scaling.

  • Recognizing homogeneity:

    • Every term in a polynomial with the same total power (e.g., x³, x²y, xy², y³) indicates homogeneity, with degree equal to that common total power.
    • Expressions of the form xᵃ yᵇ g(y/x) are homogeneous of degree a + b.
    • A function depending only on ratios, e.g., g(y/x), is degree 0.
  • Proof of Euler’s theorem (two variables): Goal: If f(λx, λy) = λⁿ f(x, y), show x f₍ₓ₎ + y f₍y₎ = n f.

    1. Define F(λ) = f(λx, λy). Then by homogeneity, F(λ) = λⁿ f(x, y).
    2. Differentiate both sides with respect to λ using chain rule: dF/dλ = f₍ₓ₎(λx, λy)·x + f₍y₎(λx, λy)·y = n λⁿ⁻¹ f(x, y).
    3. Set λ = 1: f₍ₓ₎(x, y)·x + f₍y₎(x, y)·y = n f(x, y). This is the required identity.
  • Derivation of the second-order relation (two variables): Start from Euler’s first-order identity: x z₍ₓ₎ + y z₍y₎ = n z.

    1. Differentiate both sides with respect to x: z₍ₓ₎ + x z₍ₓₓ₎ + y z₍yx₎ = n z₍ₓ₎ ⇒ x z₍ₓₓ₎ + y z₍xy₎ = (n − 1) z₍ₓ₎.
    2. Differentiate original identity with respect to y: x z₍ₓy₎ + z₍y₎ + y z₍yy₎ = n z₍y₎ ⇒ x z₍xy₎ + y z₍yy₎ = (n − 1) z₍y₎.
    3. Multiply the first of these new equations by x and the second by y, then add: x² z₍ₓₓ₎ + 2xy z₍xy₎ + y² z₍yy₎ = (n − 1)(x z₍ₓ₎ + y z₍y₎) = (n − 1)(n z) = n(n − 1) z. Assumes mixed partial derivatives are continuous so z₍xy₎ = z₍yx₎.
  • Using substitution for degree 0 cases: If u = F(y/x), set v = y/x. Then u₍ₓ₎ and u₍y₎ can be computed via chain rule:

    • v = y/x ⇒ v₍ₓ₎ = −y/x² = −v/x, v₍y₎ = 1/x.
    • u = F(v) ⇒ u₍ₓ₎ = F′(v) v₍ₓ₎, u₍y₎ = F′(v) v₍y₎. Then x u₍ₓ₎ + y u₍y₎ = x F′(v)(−v/x) + y F′(v)(1/x) = −F′(v) v + F′(v) y/x = 0, confirming degree 0 identity.
  • What to draw if asked for a diagram: Draw the xy-plane with rays from the origin. Pick a level curve of f (for example, f = constant). Show that moving along a ray (scaling x and y by λ) moves you along a curve so that the function value scales by λⁿ. Label:

    • Origin O.
    • A point P(x, y) on a ray from O.
    • A scaled point P′(λx, λy) on the same ray.
    • Indicate f(P′) = λⁿ f(P).
  • Connections to other syllabus topics (for context only): Chain rule and partial derivatives are used here (from Partial differentiation). Critical points and curvature often use the second-order identity (related to Maxima and minima of functions of two variables). Jacobians and change of variables sometimes exploit homogeneity for simplification.

Formulas

f(λx, λy) = λⁿ f(x, y) — Definition of homogeneity (degree n). Symbols: x, y are real variables (SI units depend on context); λ is a positive real scalar (dimensionless); n is real (often integer). Applies when scaling both inputs by the same factor.

f(λx₁, …, λxₖ) = λⁿ f(x₁, …, xₖ) — General k-variable version. Same meanings as above.

x·(∂z/∂x) + y·(∂z/∂y) = n·z — Euler’s theorem (two variables). z = f(x, y), degree n. Units: x and y have their own SI units; z has its unit; both sides have units of z (since ∂z/∂x has units z per unit x, multiplied by x).

Σᵢ xᵢ·(∂u/∂xᵢ) = n·u — Euler’s theorem (k variables). u = f(x₁, …, xₖ), degree n.

x²·z₍ₓₓ₎ + 2xy·z₍ₓy₎ + y²·z₍yy₎ = n(n − 1)·z — Second-order relation for C² homogeneous functions (two variables), degree n. Each term has the same units as z.

x·u₍ₓ₎ + y·u₍y₎ = 0 — Special case when u is homogeneous of degree 0 (e.g., u = F(y/x)).

If z = xᵃ yᵇ F(y/x) then degree(z) = a + b — Recognizing homogeneity in composite forms; F depends only on the ratio (degree 0).

Worked examples

Example 1 — Verify Euler’s theorem for a specific homogeneous function Given: z = (x² + y²)^(3/2). Show that x z₍ₓ₎ + y z₍y₎ = 3 z, and verify the second-order relation x² z₍ₓₓ₎ + 2xy z₍ₓy₎ + y² z₍yy₎ = 3·2 z.

Steps:

  1. Recognize degree: z = (x² + y²)^(3/2) is homogeneous of degree 3 because z(λx, λy) = (λ²x² + λ²y²)^(3/2) = (λ²)^(3/2)(x² + y²)^(3/2) = λ³ z.
  2. Compute first partials:
    • Formula: z₍ₓ₎ = ∂/∂x (x² + y²)^(3/2) = (3/2)(x² + y²)^(1/2)·2x = 3x (x² + y²)^(1/2).
    • Formula: z₍y₎ = 3y (x² + y²)^(1/2).
  3. Check Euler’s first-order identity:
    • Left side: x z₍ₓ₎ + y z₍y₎ = x[3x (x² + y²)^(1/2)] + y[3y (x² + y²)^(1/2)] = 3(x² + y²) (x² + y²)^(1/2) = 3 (x² + y²)^(3/2).
    • Right side: 3 z = 3 (x² + y²)^(3/2). They match.
  4. Compute second partials:
    • z₍ₓₓ₎ = ∂/∂x [3x (x² + y²)^(1/2)] = 3 (x² + y²)^(1/2) + 3x · (1/2)(x² + y²)^(-1/2) · 2x = 3 (x² + y²)^(1/2) + 3x² (x² + y²)^(-1/2).
    • z₍ₓy₎ = ∂/∂y [3x (x² + y²)^(1/2)] = 3x · (1/2)(x² + y²)^(-1/2) · 2y = 3xy (x² + y²)^(-1/2).
    • z₍yy₎ = 3 (x² + y²)^(1/2) + 3y² (x² + y²)^(-1/2).
  5. Verify second-order relation:
    • Left side L = x² z₍ₓₓ₎ + 2xy z₍ₓy₎ + y² z₍yy₎.
    • Substitute terms and factor (x² + y²)^(-1/2) where needed: L = x²[3 (x² + y²)^(1/2) + 3x² (x² + y²)^(-1/2)] + 2xy[3xy (x² + y²)^(-1/2)] + y²[3 (x² + y²)^(1/2) + 3y² (x² + y²)^(-1/2)]. Group as: L = 3(x² + y²)^(1/2)(x² + y²) + 3(x⁴ + 2x²y² + y⁴)(x² + y²)^(-1/2). Note x⁴ + 2x²y² + y⁴ = (x² + y²)². So L = 3(x² + y²)^(3/2) + 3(x² + y²)^(3/2) = 6 (x² + y²)^(3/2).
    • Right side R = n(n − 1) z = 3·2 (x² + y²)^(3/2) = 6 (x² + y²)^(3/2).
    • L = R, verified. Final answers: First-order identity gives 3z, second-order gives 6z.

Example 2 — Show that if u = f(y/x) then x u₍ₓ₎ + y u₍y₎ = 0 Given: u(x, y) = f(y/x). Prove Euler’s identity for degree 0 and compute u₍ₓ₎, u₍y₎ explicitly.

Steps:

  1. Let v = y/x. Then u = f(v).
  2. Compute derivatives of v:
    • v₍ₓ₎ = −y/x² = −v/x.
    • v₍y₎ = 1/x.
  3. Chain rule:
    • u₍ₓ₎ = f′(v) v₍ₓ₎ = f′(v) (−v/x).
    • u₍y₎ = f′(v) v₍y₎ = f′(v) (1/x).
  4. Form x u₍ₓ₎ + y u₍y₎:
    • x u₍ₓ₎ + y u₍y₎ = x[f′(v)(−v/x)] + y[f′(v)(1/x)] = −f′(v) v + f′(v) (y/x) = 0. Conclusion: u is homogeneous of degree 0 and satisfies the identity x u₍ₓ₎ + y u₍y₎ = 0.

Example 3 — Use Euler’s theorem to simplify a derivative calculation in three variables Given: u(x, y, z) = x² y g(y/x, z/x), where g is any differentiable function of two variables. Show u is homogeneous and compute S = x u₍ₓ₎ + y u₍y₎ + z u₍z₎ without computing all partials fully.

Steps:

  1. Recognize homogeneity: Let λ > 0.
    • u(λx, λy, λz) = (λx)² (λy) g((λy)/(λx), (λz)/(λx)) = λ³ x² y g(y/x, z/x) = λ³ u(x, y, z). So degree n = 3 (since x² y contributes degree 3; g depends only on ratios, degree 0).
  2. Apply Euler’s theorem (k = 3):
    • S = x u₍ₓ₎ + y u₍y₎ + z u₍z₎ = n u = 3 u.
  3. Conclusion without full differentiation: S = 3 x² y g(y/x, z/x).

Example 4 — Find n from Euler’s identity Given: w(x, y) = (x − 2y)⁴/(x² + y²). Find n such that x w₍ₓ₎ + y w₍y₎ = n w.

Steps:

  1. Check homogeneity by scaling λ: Numerator: (λx − 2λy)⁴ = λ⁴ (x − 2y)⁴. Denominator: (λx)² + (λy)² = λ² (x² + y²). Thus w(λx, λy) = [λ⁴ (x − 2y)⁴]/[λ² (x² + y²)] = λ² w(x, y).
  2. Therefore degree n = 2.
  3. By Euler’s theorem: x w₍ₓ₎ + y w₍y₎ = 2 w.

Common mistakes

  • Ignoring differentiability: Euler’s theorem requires at least first partial derivatives to exist. Points like (0, 0) may be excluded; state the domain.
  • Dropping units: In engineering contexts, ensure that both sides of identities have the same units; this is a good self-check.
  • Misusing the second-order relation: It applies to C² homogeneous functions; do not apply it when second derivatives are undefined or when mixed partials are not equal.
  • Errors in chain rule for ratio substitutions: When using v = y/x, students often forget v₍ₓ₎ = −v/x and v₍y₎ = 1/x.
  • Confusing “degree 0” with “zero function”: Degree 0 means scale-invariant, not identically zero.

For your exam

  • Theorems: State and prove Euler’s theorem for two variables; extend to k variables; derive the second-order identity.
  • Verifications: Given a function, verify it is homogeneous and check Euler’s identity; find its degree by scaling.
  • Applications: Use Euler’s theorem to compute x f₍ₓ₎ + y f₍y₎ (or the k-variable version) without full differentiation; simplify expressions using degree.
  • Substitution problems: For u = f(y/x), show x u₍ₓ₎ + y u₍y₎ = 0; or for u = xᵃ yᵇ F(y/x), find degree and compute Σ xᵢ u₍xᵢ₎.
  • Second-order checks: Prove or verify x² z₍ₓₓ₎ + 2xy z₍ₓy₎ + y² z₍yy₎ = n(n − 1) z for a given homogeneous function.
  • What full-marks answers include:
    • Clear statement of definitions and theorems with all assumptions (differentiability, homogeneity, domain).
    • Step-by-step derivations using chain rule and careful differentiation.
    • Explicit identification of the degree using f(λx, …) = λⁿ f(x, …).
    • If asked, a neat diagram showing scaling along rays from the origin, with labels.
    • Clean algebra, correct handling of partial derivatives, and a final boxed or highlighted identity.
    • For verification problems, show both sides match exactly, not just numerically at a point.

Quick check

  1. If f(x, y) is homogeneous of degree n, what is f(3x, 3y) in terms of f(x, y)?
  2. Let u(x, y) = y/x + x/y. Is u homogeneous? If yes, what is its degree, and what is x u₍ₓ₎ + y u₍y₎?
  3. Suppose w(x, y) = x² y³ h(y/x) where h is differentiable. Find the degree of w and the value of x w₍ₓ₎ + y w₍y₎ in terms of w.

Answers: 1) f(3x, 3y) = 3ⁿ f(x, y). 2) Yes, degree 0 (since each term is a function of y/x); x u₍ₓ₎ + y u₍y₎ = 0. 3) Degree 5; x w₍ₓ₎ + y w₍y₎ = 5 w.

Finished this topic? Mark it so your progress, study plan and readiness keep up.

Stuck on something here?