Binomial formula for Macdonald polynomials and applications (Q1372891)

The author gives the non-evident and the very non-trivial generalization of the Macdonald polynomials. He defines the ``interpolation Macdonald polynomial'' \(P^*(x_1,\dots, x_n;q,t)\) (\(q\) and \(t\) are two parameters) by the following conditions: (1) \(P^*\) is symmetric in variables \(x_i t_i^{-1}\), (2) \(\deg P_\mu^*=| \mu|\) for any partition \(\mu\) (\(\mu\) has \(\leq n\) parts), (3) \(P_\mu^* (q^\lambda;q,t)= 0\) for all partitions \(\lambda\) such that \(\mu\not\subset \lambda\). With the additional condition of the normalization this polynomial is unique. For these polynomials the author gives the explicit binomial theorem (the explicit expression of \(P_\lambda^* (ax_1,\dots, ax_n;q,t)\) in terms of \(P_\mu^* (x_1,\dots, x_n; \frac 1q, \frac 1t)\), \(\mu\subset \lambda\)). There is the explicit difference equation for \(P_\mu^*\) and there is the efficient algorithm for multivariate symmetric Newton interpolation.