[Paper Review] Littelmann path model for geometric crystals

Reda Chhaibi|arXiv (Cornell University)|May 25, 2014

Algebraic structures and combinatorial models18 references3 citations

TL;DR

This paper constructs a geometric Littelmann path model for geometric crystals in the sense of Berenstein and Kazhdan, using continuous paths in the real Cartan subalgebra 𝔞, with root operators defined via integral transforms that geometrically lift Littelmann’s piecewise-linear reflections. The key result is a crystal morphism from the path model to the group picture that restricts to an isomorphism on connected components, realizing the geometric Robinson-Schensted correspondence and identifying the geometric Pitman transform as the highest weight path transform.

ABSTRACT

We construct a path model for geometric crystals in the sense of Berenstein and Kazhdan. Our model is in every way similar to Littelmann's and tropicalizes to his path model. This paper lays the foundational material for a subsequent work where we examine the measure induced on geometric crystals by Brownian motion. If we call Berenstein and Kazhdan's realization of geometric crystals the group picture, we prove that the path model projects onto the group picture thanks to a morphism of crystals that restricts to an isomorphism on connected components. This projection is in fact the geometric analogue of the Robinson-Schensted correspondence and involves solving a left-invariant differential equation on the Borel subgroup. Moreover, we identify the geometric Pitman transform $\mathcal{T}_{w_0}$ introduced by Biane, Bougerol and O'Connell as the transform giving the path with highest weight, in the geometric crystal path model. This allows to prove a geometric version of Littelmann's independence theorem. The geometric Robinson-Schensted correspondence is detailed in a special section, because of its importance. Finally, we exhibit the Kashiwara and Schützenberger involutions in both the group picture and the path model. In an appendix, we explain how the left-invariant flow is related to the image of the Casimir element in Kostant's Whittaker model.

Motivation & Objective

• To develop a geometric analogue of Littelmann’s path model for geometric crystals, extending his combinatorial framework to the geometric setting.
• To establish a morphism from the path model to Berenstein and Kazhdan’s group picture of geometric crystals, proving it is an isomorphism on connected components.
• To identify the geometric Pitman transform as the transform yielding the highest weight path in the path model, thereby realizing a geometric version of the Robinson-Schensted correspondence.
• To exhibit the Kashiwara and Schützenberger involutions in both the group picture and the path model, providing geometric lifts of these classical crystal operations.
• To connect the left-invariant differential equation on the Borel subgroup to Kostant’s Whittaker model and the quantum Toda Hamiltonian, linking geometric crystals to stochastic processes and integrable systems.

Proposed method

• Constructs a geometric crystal structure on the space of continuous paths in the real Cartan subalgebra 𝔞, using rescaling to generate a family of equivalent models.
• Defines root operators as integral transforms that geometrically lift Littelmann’s piecewise-linear reflection scheme, preserving the crystal structure.
• Establishes a morphism from the path model to the group picture (totally positive elements in the Borel subgroup B), showing it restricts to an isomorphism on connected components.
• Identifies the geometric Pitman transform 𝒯_{w₀} as the transform that maps any path to its highest weight representative in the path model.
• Uses string and Lusztig parameters to parametrize paths, with inversion lemmas enabling coordinate-based crystal actions and connectedness criteria.
• Relates the left-invariant flow on B to the image of the Casimir element in Kostant’s Whittaker model, showing it reduces to the quantum Toda Hamiltonian on 𝔞.

Experimental results

Research questions

• RQ1 How can Littelmann’s path model be geometrically lifted to the setting of geometric crystals in the sense of Berenstein and Kazhdan?
• RQ2 What is the geometric analogue of the Robinson-Schensted correspondence, and how is it realized via path transforms?
• RQ3 How does the geometric Pitman transform relate to the highest weight path in the path model?
• RQ4 How are the Kashiwara and Schützenberger involutions realized in both the group picture and the path model?
• RQ5 What is the connection between the left-invariant differential equation on B and the quantum Toda Hamiltonian in Kostant’s Whittaker model?

Key findings

• The path model projects onto the group picture via a crystal morphism that restricts to an isomorphism on connected components, establishing a geometric lift of the crystal isomorphism.
• The geometric Pitman transform 𝒯_{w₀} is identified as the transform that yields the highest weight path in the path model, realizing the geometric Robinson-Schensted correspondence.
• The path model is shown to be isomorphic to the geometric crystal of continuous paths in the real Cartan subalgebra 𝔞, with tensor products given by path concatenation and weight maps by path endpoints.
• The left-invariant differential equation on the Borel subgroup B is shown to be governed by the operator Ω_χ, which reduces to the quantum Toda Hamiltonian on 𝔞, linking geometric crystals to integrable systems.
• The Kashiwara and Schützenberger involutions are explicitly constructed in both the group picture and the path model, providing geometric lifts of these fundamental crystal operations.
• The geometric Littelmann model tropicalizes to Littelmann’s original path model, confirming consistency with the classical theory.