Suppose first that is orthogonal to . Every can be written with . The Pythagorean theorem in an inner-product space gives
Thus is a best approximation.
Conversely, if minimizes the distance, then for every the quadratic
has its minimum at . Differentiating there gives in the real case; varying real and imaginary parts gives the complex case. Therefore
The defining relation for gives
Part (a) therefore shows that is the best approximation. Moreover, and are orthogonal, so
Thus orthogonal projection is a contraction:
The integral form of the Taylor theorem about the left endpoint is
where . Apply the order- divided difference at . The polynomial term vanishes, and linearity permits interchange with the integral:
By the definition ,
This is the Peano kernel theorem for the divided-difference functional.
Substitute the B-spline expansion into . The coefficients solve
with the Gram matrix and right-hand side
The linearly independent B-splines have a positive-definite Gram matrix, so this system has a unique coefficient vector.
For any admissible , part (a) rewrites the constraints as
Hence is orthogonal to every and therefore to their span, which contains . The Pythagorean theorem in an inner-product space gives
Choose any -fold antiderivative of . The identity from part (a) and show that satisfies all prescribed divided differences, and equality holds in the norm bound. Therefore
characterizes the minimizers. They are unique up to addition of an arbitrary polynomial in , which changes neither the order- divided differences nor the th derivative.

Articles by others on the same topic (0)

There are currently no matching articles.