The function has weak derivative and the required endpoint values, so it belongs to the admissible Sobolev space. Its fourth power is , and hence
Since the integrand is nonnegative, zero is a global lower bound. Thus is a global minimizer. Its derivative does not belong to , since its power is proportional to ; positive regularization therefore needs different competitors.
A convenient form of the Tonelli existence theorem is: on a finite interval with prescribed endpoint values, if a continuous integrand is convex in and satisfies for some , then, provided a finite-energy admissible competitor exists, its integral attains a minimum in the corresponding class. The lower bound may more generally subtract an integrable function of . Compact one-dimensional Sobolev embedding controls , while convexity gives weak lower semicontinuity in the derivative.
Here and is continuous, convex in , and bounded below by . The competitor has finite energy. Thus for every the Tonelli existence theorem gives an admissible minimizer in .
Any finite-energy member of the larger stated class necessarily has derivative in and bounded function values by absolute continuity; it therefore belongs to . All other competitors have infinite energy. Hence the same function is a global minimizer in the full stated class.
The finite-energy comparison gives . Together with the endpoint value and absolute continuity, this proves .
For , the power competitor is admissible and has finite derivative energy. Direct integration gives
Choose and . The first three terms tend to zero as , while the last equals . By minimality and nonnegativity,
One also obtains the quantitative bound : for , the mean value theorem gives , and . Thus the squared mismatch is at most , matching the order of the derivative penalty with this choice. This is a power-law comparison for vanishing derivative regularization.

Articles by others on the same topic (0)

There are currently no matching articles.