Let be a smooth constrained minimizer. For any smooth variation vanishing at the endpoints, use the normalized admissible curve . Since , differentiating at zero givesIn particular this holds for every compactly supported test function in . Integration by parts and the fundamental lemma of the calculus of variations givewhich is the required Bessel differential equation with . The lemma used here says that a continuous function whose integral against every compactly supported smooth test function is zero must itself vanish. No multiplier-existence theorem is needed: the explicit normalization produces the variation directly.
The endpoint causes no divergent energy for these variations: smoothness and imply . Also for every nonzero admissible , and has unit and energy . Thus minimizing with is equivalent to minimizing the scale-invariant Rayleigh quotient . The zero function is excluded whenever this quotient is written.
Set and . The chain rule turns the equation into the parameter-one Bessel differential equation. By the regular-at-zero solution property supplied in the paper,For a nonzero solution, , so forces . Conversely, every positive zero supplies the admissible eigenfunction with eigenvalue .
Here is a direct justification of the infimum characterization that does not assume a minimizer exists. Let and choose its sign so that on . The first positive zero exists by the oscillatory large-argument behavior of the Bessel function of the first kind. At zero, its Frobenius series for Bessel functions gives . At one, its zero is simple by uniqueness for a regular second-order ordinary differential equation. Expanding and using the equation for gives the Ground-state representation of a Sturm-Liouville energy:For completeness, apply the identity first on . The discarded boundary term is . It tends to zero at both ends: and at zero, while and at one. Taking achieves equality, soThe infimum is over the stated smooth admissible functions with finite energy; the equality also proves attainment by a multiple of .
The trial function givesIts Rayleigh quotient is , yielding the upper bound. Conversely, on and the derivative energy is nonnegative, so for every admissible . Part (b) therefore gives
Articles by others on the same topic
There are currently no matching articles.