The theory concerned with finding extrema (minima, maxima, stationary points) of functionals, typically integrals depending on functions and their derivatives, and deriving associated Euler–Lagrange equations and variational inequalities.

Principle

Translate optimization over functions into differential conditions: compute first variations to obtain Euler–Lagrange equations, use second variation or convexity for stability, and apply direct methods (lower semicontinuity, coercivity) to prove existence in appropriate function spaces (e.g., Sobolev spaces).

Demonstration

Derive the Euler–Lagrange equation for the functional ∫_a^b L(x,u,u') dx to obtain boundary value problems; solve the brachistochrone or minimal surface problem as classical examples; use the Rayleigh quotient min_{u≠0} (⟨Au,u⟩/⟨u,u⟩) to characterize eigenvalues variationally.

Misapplication

Varying boundary values without enforcing admissible variations, or applying Euler–Lagrange formal manipulations when the minimizer lies outside the smooth class (necessitating weak formulations and Sobolev spaces), leads to incorrect conclusions about existence or necessary conditions.

Consequence

Proper variational methods yield existence of minimizers, derive PDEs as Euler–Lagrange equations, provide stability criteria, and underpin numerical approximation schemes (finite elements) and physical principles (least action).

Reversal

Invert to pointwise optimization where one optimizes values at each point independently: this neglects coupling through derivatives and boundary conditions, destroying the global variational structure and often failing to capture PDE constraints.

Boundary

Deals with functionals on spaces of functions, typically requiring Sobolev or other function-space settings for lower semicontinuity and compactness; excludes discrete optimization problems lacking a variational integral structure and purely algebraic minimization without functional dependence on derivatives.

Semantic Tension

Tension exists between classical smooth variational calculus and weak variational formulations: notions of admissible variations, regularity, and appropriate function spaces can conflict; similarly, optimization formulations that are not coercive or lower semicontinuous require different techniques.

Synthesis

Calculus of variations frames extremal problems for functionals by converting variations into differential (Euler–Lagrange) conditions and employing direct existence methods and regularity theory; it connects geometry, PDEs, mechanics, and numerical methods by identifying when functionals admit minimizers and how those minimizers satisfy differential constraints.