Nijenhuis tensor and invariant polynomials (Q2169846)

The notion of symplectic groupoid was introduced by \textit{A. Weinstein} [Bull. Am. Math. Soc., New Ser. 16, 101--104 (1987; Zbl 0618.58020)]. The basic idea was that a proper quantization must be compatible with the additional groupoid structure. If the quantization scheme is given by geometric quantization, both prequantization and polarization should be compatible with the groupoid structure. Furthermore, if the groupoid is prequantizable as a symplectic manifold, then such compatible prequantization always exists and is unique. A nondegenerate symplectic Poisson-Nijenhuis structures (PN) lead to multiplicative polarizations. A symplectic PN structure on a smooth manifold \(M\) consists of a symplectic structure \(\omega\) and a Poisson structure \(\pi\) such that some compatibility condition holds. Motivated by the problem of quantization, the study of these PN geometries was started in [\textit{F. Bonechi} et al., J. Symplectic Geom. 16, No. 5, 1167--1208 (2019; Zbl 1422.53068)], where the diagonalization of the Nijenhuis tensor was solved for the classical cases (AIII, BI, DIII, CI). In this paper the authors discuss the diagonalization problem of the Nijenhuis tensor in a class of Poisson-Nijenhuis structures defined on compact Hermitian symmetric spaces. They study its action on the ring of invariant polynomials of a Thimm chain of subalgebras. The existence of \(\phi\)-minimal representations defines a suitable basis of invariant polynomials that completely solves the diagonalization problem. Such representations exist in the classical cases AIII, BDI, DIII and CI, and do not exist in the exceptional cases EIII and EVII.