Fundamental Lemma Of Calculus Of Variations

The Fundamental Lemma of Calculus of Variations: The Hidden Engine of Optimization

At the heart of the calculus of variations lies a deceptively simple yet profoundly powerful statement known as the fundamental lemma of calculus of variations. This lemma is not merely a technical tool; it is the critical bridge that transforms the abstract problem of finding a function that minimizes an integral into a concrete, solvable differential equation. It provides the rigorous justification for a step that feels almost like magic: concluding that if the integral of a function multiplied by any smooth test function is zero, then the original function itself must be zero. This principle is the silent workhorse behind the derivation of the Euler-Lagrange equation, the cornerstone of fields from classical mechanics to modern geometric optics.

What is the Fundamental Lemma?

In its most common form, the lemma addresses a specific scenario. Consider a function ( f(x) ) defined on an interval ([a, b]). The lemma states:

If the integral of ( f(x) ) multiplied by an arbitrary smooth function ( \eta(x) ) that vanishes at the endpoints is zero, then ( f(x) ) must be identically zero on ([a, b]).

More formally: Let ( f ) be a continuous function on ([a, b]). If [ \int_a^b f(x) \eta(x) , dx = 0 ] for all functions ( \eta(x) ) that are continuously differentiable (( C^1 )) and satisfy ( \eta(a) = \eta(b) = 0 ), then ( f(x) = 0 ) for all ( x \in [a, b] ).

The power of the lemma comes from the word "arbitrary." The test function ( \eta(x) ) is not a specific function; it is any function from a broad, well-defined class. The condition that ( \eta(a) = \eta(b) = 0 ) is crucial—it corresponds to the endpoints being fixed in the original variational problem. The lemma essentially says that the only function "orthogonal" to all such test functions is the zero function itself.

Why Is This Lemma So Important?

The calculus of variations seeks to minimize (or maximize) a functional, which is a function of a function. A classic example is finding the curve ( y(x) ) that minimizes the arc length or the action in physics. The standard procedure is:

1. Consider a variation ( y(x) + \epsilon \eta(x) ), where ( \eta(x) ) is a small, smooth perturbation that is zero at the endpoints (since the endpoints are usually fixed).
2. Compute the first variation of the functional, ( \delta J ), which is the derivative with respect to ( \epsilon ) at ( \epsilon = 0 ). This yields an expression of the form: [ \delta J = \int_a^b F(x, y, y') \eta(x) , dx ] where ( F ) is the integrand of the original functional.
3. For ( y(x) ) to be an extremal (minimizer or maximizer), this first variation must be zero for all admissible ( \eta(x) ).
4. Here is the critical step: The expression ( \int_a^b (\text{some expression}) \eta(x) , dx = 0 ) for all ( \eta(x) ). By the fundamental lemma, if the "some expression" is continuous, we can immediately conclude that the "some expression" itself must be identically zero.

This final conclusion is precisely the Euler-Lagrange equation: [ \frac{\partial F}{\partial y} - \frac{d}{dx}\left(\frac{\partial F}{\partial y'}\right) = 0. ] Without the fundamental lemma, we would be stuck with an integral condition that holds for many functions, unable to deduce the powerful pointwise differential equation that is often much easier to solve.

A Detailed Look at the Proof

Understanding the proof solidifies why the lemma's conditions are necessary. The standard proof for a continuous ( f(x) ) proceeds by contradiction and clever construction.

Proof Sketch:

1. Assume ( f(x) ) is not identically zero. Then there exists some point ( c \in (a, b) ) where ( f(c) \neq 0 ). Without loss of generality, assume ( f(c) > 0 ).
2. Use continuity. Since ( f ) is continuous, there is a small interval ([c - \delta, c + \delta]) around ( c ) where ( f(x) > \alpha > 0 ) for some positive constant ( \alpha ).
3. Construct a specific test function ( \eta(x) ). We need a ( C^1 ) function that is:
   • Zero outside ([c - \delta, c + \delta]).
   • Zero exactly at ( x = c - \delta ) and ( x = c + \delta ).
   • Positive on the open interval ((c - \delta, c + \delta)). A classic choice is a "bump function" or a piecewise quadratic function that smoothly rises from zero at ( c-\delta ) to a peak at ( c ) and falls back to