Skip to content

Calculus of Variations Formula Sheet and Short Notes

One-page CSIR NET calculus of variations revision: Euler–Lagrange and its extensions, Legendre, Jacobi and Weierstrass tests, Ritz, Galerkin, Rayleigh.

  • 7 min read
  • Updated
On this page · 6 sections
  1. The paper in one table
  2. Variation and the Euler–Lagrange equation
  3. Extensions
  4. Conditions for extrema
  5. Variational methods for BVPs
  6. Last-minute checklist

Every formula, condition and counter-example from Functionals, First Variation and the Euler–Lagrange Equation, Extensions of the Euler–Lagrange Equation, Conditions for Extrema and Variational Methods for Boundary Value Problems, on one page.

The paper in one table

Per the NTA information bulletin for December 2026 (Mathematical Sciences, code 704):

PartGivenAttemptCorrectWrongMaximum
B4025+3−0.7575
C6020+4.75095

Part C: one or more options may be correct; marks only if you mark all correct options and no wrong one. No partial credit, no negative marking. Details on the Mathematical Sciences overview.

Variation and the Euler–Lagrange equation

$$J[y] = \int_a^b F(x, y, y')\, dx, \qquad \delta J = \frac{d}{d\varepsilon}J[y + \varepsilon\eta]\Big|_{\varepsilon = 0} = \int_a^b \Big(F_y - \frac{d}{dx}F_{y'}\Big)\eta\, dx + \Big[F_{y'}\eta\Big]_a^b$$$$\boxed{F_y - \frac{d}{dx}F_{y'} = 0} \qquad \text{(second order when } F_{y'y'} \ne 0\text{)}$$
SituationShortcut
no \( y \) in \( F \)\( F_{y'} = C \)
no \( x \) in \( F \)\( F - y'F_{y'} = C \) (Beltrami; check constant solutions separately)
\( F = F(y') \)straight lines
\( F = F(x, y) \)\( F_y = 0 \), algebraic, usually no extremal
\( F = M + Ny' \), \( M_y = N_x \)\( J \) path-independent: every curve is an extremal
\( y(b) \) freenatural condition \( F_{y'}(b) = 0 \)
IntegrandExtremals
\( \sqrt{1 + y'^2} \)lines
\( y\sqrt{1 + y'^2} \)catenaries \( c\cosh\frac{x - x_0}{c} \)
\( \frac{\sqrt{1 + y'^2}}{\sqrt{y}} \)cycloids (brachistochrone)
\( \frac{\sqrt{1 + y'^2}}{y} \)circles centred on the \( x \)-axis
\( y'^2 \pm y^2 \)\( y'' = \pm y \): \( \cosh, \sinh \) or \( \cos, \sin \)
\( y'^2 + 2f(x)y \)\( y'' = f \)
\( x^2y'^2 + ky^2 \)Euler–Cauchy \( x^r \), \( r^2 + r - k = 0 \)

Extensions

ProblemEquationConditions / constants
\( F(x, y, \dots, y^{(n)}) \)\( \sum_{k=0}^{n} (-1)^k\frac{d^k}{dx^k}F_{y^{(k)}} = 0 \)order \( 2n \); \( y, \dots, y^{(n-1)} \) at both ends
\( n \) functions \( y_i \)\( F_{y_i} - \frac{d}{dx}F_{y_i'} = 0 \) each\( 2n \) constants
\( u(x, y) \)\( F_u - \partial_xF_{u_x} - \partial_yF_{u_y} = 0 \)PDE (Euler–Ostrogradsky)
\( \int G = \ell \)E–L of \( F + \lambda G \), \( \lambda \) constant\( c_1, c_2, \lambda \) from ends + constraint
\( g(x, y, z) = 0 \)\( F + \lambda(x)g \)\( \lambda \) a function
end on \( y = \varphi(x) \)\( F + (\varphi' - y')F_{y'} = 0 \)for \( f\,ds \): orthogonality \( y'\varphi' = -1 \)
end on \( y = y_1 \), \( x_1 \) free\( F - y'F_{y'} = 0 \)
mechanics\( \delta\int (T - V)\, dt = 0 \)Lagrange’s equations
Integrand in two variablesEquation
\( u_x^2 + u_y^2 \)Laplace \( \Delta u = 0 \)
\( u_x^2 + u_y^2 + 2fu \)Poisson \( \Delta u = f \)
\( u_x^2 + u_y^2 + u^2 \)\( \Delta u = u \)
\( \frac{1}{2}(u_t^2 - c^2u_x^2) \)wave \( u_{tt} = c^2u_{xx} \)
\( \sqrt{1 + u_x^2 + u_y^2} \)minimal surface equation
  • Dido: max area for fixed length gives circular arcs.
  • Eigenvalues: min of \( \int (py'^2 + qy^2) \) with \( \int ry^2 = 1 \) gives \( -(py')' + qy = \lambda ry \); the multiplier is the eigenvalue.

Conditions for extrema

$$\delta^2J[h] = \int_a^b (Ph'^2 + Qh^2)\, dx, \qquad P = F_{y'y'}, \quad Q = F_{yy} - \frac{d}{dx}F_{yy'}$$$$\text{Jacobi: } -(Pu')' + Qu = 0, \quad u(a) = 0,\ u'(a) = 1; \qquad E(x, y, p, w) = F(x, y, w) - F(x, y, p) - (w - p)F_{y'}(x, y, p)$$
ConditionFor a minimumStatus
Euler–Lagrange\( F_y - \frac{d}{dx}F_{y'} = 0 \)necessary
Legendre\( F_{y'y'} \ge 0 \)necessary
Jacobino conjugate point in \( (a, b) \)necessary (\( P > 0 \))
Weierstrass\( E \ge 0 \) for all \( w \) along the extremalnecessary for strong
E–L + \( F_{y'y'} > 0 \) + no conjugate point in \( (a, b] \)sufficient for weak
+ field of extremals + \( E \ge 0 \) nearbysufficient for strong
\( F \) convex in \( (y, y') \)Hessian PSDextremal is the global min
corners\( F_{y'} \), \( F - y'F_{y'} \) continuousWeierstrass–Erdmann

Maximum: reverse Legendre and Weierstrass; Jacobi unchanged. Strong \( \Rightarrow \) weak.

Counter-exampleLesson
\( \int_0^b (y'^2 - y^2) \), zero ends\( y = 0 \) strong min for \( b < \pi \); not a min for \( b > \pi \) (conjugate point \( \pi \)); non-unique at \( b = \pi \). Legendre is not sufficient
\( y'^2 - k^2y^2 \)conjugate point \( \frac{\pi}{k} \)
\( \int y'^3 \), line of slope \( p > 0 \)weak min, not strong: \( E = (w - p)^2(w + 2p) \)
\( \int (y'^2 - 1)^2 \)broken extremals (slopes \( \pm 1 \)) beat the smooth one; minimiser not unique
\( \int_{-1}^1 x^2y'^2 \), \( y(\pm 1) = \pm 1 \)\( \inf = 0 \), not attained
\( \int y\sqrt{1 + y'^2} \), ends far apartno catenary fits the boundary values

Variational methods for BVPs

$$\boxed{-(py')' + qy = f \;\longleftrightarrow\; J[y] = \int_a^b (py'^2 + qy^2 - 2fy)\, dx}$$
FactDetail
Self-adjoint formmultiply \( y'' + Py' + Qy = R \) by \( \mu = e^{\int P} \)
Uniqueness\( p > 0 \), \( q \ge 0 \): convex, unique minimiser
Essential conditionsDirichlet: imposed on trial functions
Natural conditionsNeumann \( py' = 0 \); Robin via \( +\sigma y(b)^2 \): not imposed
Ritzminimise \( J(c_1, \dots, c_n) \); linear symmetric system; \( J[y_n] \ge \min J \), decreasing in \( n \)
Galerkin\( \int R\phi_k = 0 \); equals Ritz for self-adjoint problems; works without a functional
Collocation\( R(x_k) = 0 \)
\( J_{\min} \)\( -\int fy \) at the true solution
Rayleigh\( \lambda_1 = \min \frac{\int (py'^2 + qy^2)}{\int ry^2} \); every admissible trial is an upper bound
Dirichlet principleharmonic \( u \) minimises \( \iint \lvert\nabla u\rvert^2 \) for given boundary values
Poisson \( -\Delta u = f \)minimise \( \iint (\lvert\nabla u\rvert^2 - 2fu) \)

Numbers worth remembering (all on \( [0, 1] \) unless stated):

ItemValue
\( \int (1 - 2x)^2 \), \( \int x(1 - x) \), \( \int x^2(1 - x)^2 \)\( \frac{1}{3} \), \( \frac{1}{6} \), \( \frac{1}{30} \)
\( y'' + y + x = 0 \), one-term Ritz \( c\,x(1 - x) \)\( c = \frac{5}{18} \); \( y(\frac{1}{2}) \approx 0.0694 \) vs exact 0.0698
\( -y'' + y = x \), one-term Ritz\( c = \frac{5}{22} \)
\( -y'' = \lambda y \), Dirichlet, trial \( x(1 - x) \)\( 10 \) vs \( \pi^2 \approx 9.87 \)
\( -y'' = \lambda y \), \( y(0) = 0 \), \( y'(1) = 0 \), trials \( x \), \( 2x - x^2 \)\( 3 \), \( \frac{5}{2} \) vs \( \frac{\pi^2}{4} \approx 2.47 \)
\( -\Delta u = 1 \), unit square, trial \( c\,x(1 - x)y(1 - y) \)\( c = \frac{5}{4} \); centre \( \frac{5}{64} \) vs about 0.0737
\( -\Delta u = \lambda u \), unit square, same trial\( 20 \) vs \( 2\pi^2 \approx 19.74 \)

Last-minute checklist

Back to the Calculus of Variations chapter overview, or revise its foundations in Real Analysis: Quick Revision.

Advertisement
Advertisement