English

From Darboux-Egorov system to bi-flat $F$-manifolds

Mathematical Physics 2015-06-05 v3 math.MP Exactly Solvable and Integrable Systems

Abstract

Motivated by the theory of integrable PDEs of hydrodynamic type and by the generalization of Dubrovin's duality in the framework of FF-manifolds due to Manin [22], we consider a special class of FF-manifolds, called bi-flat FF-manifolds. A bi-flat FF-manifold is given by the following data (M,1,2,,,e,E)(M, \nabla_1,\nabla_2,\circ,*,e,E), where (M,)(M, \circ) is an FF-manifold, ee is the identity of the product \circ, 1\nabla_1 is a flat connection compatible with \circ and satisfying 1e=0\nabla_1 e=0, while EE is an eventual identity giving rise to the dual product *, and 2\nabla_2 is a flat connection compatible with * and satisfying 2E=0\nabla_2 E=0. Moreover, the two connections 1\nabla_1 and 2\nabla_2 are required to be hydrodynamically almost equivalent in the sense specified in [2]. First we show that, similarly to the way in which Frobenius manifolds are constructed starting from Darboux-Egorov systems, also bi-flat FF-manifolds can be built from solutions of suitably augmented Darboux-Egorov systems, essentially dropping the requirement that the rotation coefficients are symmetric. Although any Frobenius manifold possesses automatically the structure of a bi-flat FF-manifold, we show that the latter is a strictly larger class. In particular we study in some detail bi-flat FF-manifolds in dimensions n=2, 3. For instance, we show that in dimension 3 bi-flat FF-manifolds are parametrized by solutions of a two parameters Painlev\'e VI equation, admitting among its solutions hypergeometric functions. Finally we comment on some open problems of wide scope related to bi-flat FF-manifolds.

Keywords

Cite

@article{arxiv.1205.2468,
  title  = {From Darboux-Egorov system to bi-flat $F$-manifolds},
  author = {Alessandro Arsie and Paolo Lorenzoni},
  journal= {arXiv preprint arXiv:1205.2468},
  year   = {2015}
}

Comments

32 pages, eliminated a remark at the end of proof of Theorem 6.2