SummaryWe try to solve the bivariate interpolation problem (1.3) for polynomials (1.1), whereS is a lower set of lattice points, and for theq-th interpolation knot,Aq is the set of orders of derivatives that appear in (1.3). The number of coefficients |S| is equal to the number of equations Σ|Aq|. If this is possible for all knots in general position, the problem is almost always solvable (=a.a.s.). We seek to determine whether (1.3) is a.a.s. An algorithm is given which often gives a positive answer to this. It can be applied to the solution of a problem of Hirschowitz in Algebraic Geometry. We prove that for Hermite conditions (1.3) (when allAq are lower triangles of orderp) andP is of total degreen, (1.3) is a.a.s. for allp=1, 2, 3 and alln, except for the two casesp=1,n=2 andp=1,n=4.
[1]
George G. Lorentz,et al.
Solvability problems of bivariate interpolation I
,
1986
.
[2]
H. Hakopian.
Multivariate divided differences and multivariate interpolation of Lagrange and Hermite type
,
1982
.
[3]
Charles A Micchelli,et al.
A Constructive Approach to Kergin Interpolation in R(k).
,
1978
.
[4]
R. Lorentz.
Uniform bivariate Hermite interpolation
,
1990
.
[5]
A. Hirschowitz.
La Methode d1Horace pour l'Interpolation à Plusieurs Variables
,
1985
.
[6]
Solvability of multivariate interpolation
,
1989
.
[7]
L. Mordell,et al.
Diophantine equations
,
1969
.