The global extension problem, crossed products and co-flag non-commutative Poisson algebras

TL;DR

This paper classifies all Poisson algebra structures on a vector space E that extend a given Poisson algebra P, using a cohomological framework and providing explicit examples including co-flag Poisson algebras.

Contribution

It introduces a comprehensive classification method for Poisson algebra extensions via a new cohomological set, generalizing classical cohomology and applying to complex algebra structures.

Findings

Classified Poisson algebra structures using a cohomological set ${ m extbf{GP} extbf{H}}^{2}(P,V)$.

Connected classical cohomology group $H^2(P,V)$ as a fundamental component.

Provided explicit examples for metabelian and co-flag Poisson algebras.

Abstract

Let $P$ be a Poisson algebra, $E$ a vector space and $\pi : E \to P$ an epimorphism of vector spaces with $V = {\rm Ker} (\pi)$. The global extension problem asks for the classification of all Poisson algebra structures that can be defined on $E$ such that $\pi : E \to P$ becomes a morphism of Poisson algebras. From a geometrical point of view it means to decompose this groupoid into connected components and to indicate a point in each such component. All such Poisson algebra structures on $E$ are classified by an explicitly constructed classifying set ${\mathcal G} {\mathcal P} {\mathcal H}^{2} \, (P, \, V)$ which is the coproduct of all non-abelian cohomological objects ${\mathcal P} {\mathcal H}^{2} \, (P, \, (V, \cdot_V, [-,-]_V))$ which are the classifying sets for all extensions of $P$ by $(V, \cdot_V, [-,-]_V)$. The second classical Poisson cohomology group $H^2 (P, V)$ appears as the most elementary piece among all components of ${\mathcal G} {\mathcal P} {\mathcal H}^{2} \, (P, \, V)$. Several examples are provided in the case of metabelian Poisson algebras or co-flag Poisson algebras over $P$: the latter being Poisson algebras $Q$ which admit a finite chain of epimorphisms of Poisson algebras $P_n : = Q \stackrel{\pi_{n}}{\longrightarrow} P_{n-1} \, \cdots \, P_1 \stackrel{\pi_{1}} {\longrightarrow} P_{0} := P$ such that ${\rm dim} ( {\rm Ker} (\pi_{i}) ) = 1$, for all $i = 1, \cdots, n$.