Constrained degree reduction of polynomials in Bernstein-Bézier form over simplex domain (Q2483332)

Degree reduction of Bézier curves belongs to the important problems of computer aided geometric design. The degree reduction strategies can be classified by different norms in which the distance between Bernstein polynomials is measured. In this paper, the constrained degree reduction of a polynomial of degree \(n\) over the \(d\)-dimensional simplex domain in \(L_2\)-norm is presented. First, the authors show that the orthogonal complements of a subspace in the constrained polynomial space of degree \(n\) over the simplex domain with respect to the \(L_2\)-inner product and the weighted Euclidean inner product of the Bernstein-Bézier coefficients are equal for some weights which are proposed in the paper. After that, the proof that the best constrained degree reduction of a polynomial of degree \(n\) over the simplex domain in \(L_2\)-norm is equal to the best approximation of the vector of coefficients from all vectors of coefficients of degree elevated polynomials with the constraint in weighted the Euclidean norm of vectors is given.