Free Lie algebroids and the space of paths (Q2474169)

The goal of the article is to construct algebraic and algebro-geometric models for spaces of paths. Paths in a manifold \(X\), considered up to reparametrizations and cancelations, form a groupoid \(\Pi_X.\) The author constructs a Lie algebroid \(\mathcal{P}_X\) which play the role of the Lie algebra for \(\Pi_X,\) i.e. describes infinitesimal paths. When \(X\) is an algebraic variety, an algebro-geometric model for the formal neighborhood of \(X\) in \(\Pi_X\) is constructed. This is an ind-scheme, denoted \(\widehat\Pi_X\) which also have a groupoid structure. The article applies the Kontsevich moduli stack \(M_{g,2}(X,\beta)\) of stable maps from 2-pointed curves \((C,x,y)\) of genus \(g\) into \(X\) which is related to the space of unparametrized paths. One of the main results of the article constructs a morphism from a certain formal neighborhood in \(M_{0,2}(X,\beta)\) into \(\widehat\Pi_X\) in the \(g=0\) case. Any vector bundle with connection on \(X\) gives a representation of \(\Pi_X\) by holonomy. Representations of \(\Pi_X\) are the same as connections so that that the Lie algebroid \(\mathcal{P}_X\) can be seen as the universal receptacle of the curvature data for all connections. It appears that \(\mathcal{P}_X\) is a fundamental differential-geometric object associated to any manifold. Another main result in the article identifies formal germs of connections on \(X\) at \(x\) with representations of a certain infinite-dimensional Lie algebra \(\mathcal{P}(X,x)\) which the author calls the \textit{fundamental Lie algebra}. This provides a Taylor formula for connections. The author's construction of \(\mathcal{P}_X\) is also a version of the free Lie algebra construction but in the context of Lie algebroids. In the curvature language the appearance of the (free) Lie algebras here reflects the interpretation of the Bianchi identity as the Jacobi identity for the curvature operators \(\nabla_i=\delta_i+A_i\). Sections of \(\mathcal{P}_X\) give rise to ``noncommutative vector fields'', i.e. to natural systems of operators \(\{P_{E,\nabla}:E\to E\}\) defined for all bundles with connections \((E,\nabla)\) on \(X\) and satisfying the Leibniz rule with respect to the tensor product. By applying the enveloping algebra construction the result is a sheaf of rings \(\mathbb{D}_X=U(\mathcal{P}_X)\) whose sections are called noncommutative differential operators. Sections of \(\mathbb{D}_X\) are identified with these operators satisfying an extra assumption of regularity. It is then interesting to study natural systems of pseudo-differential operators because intuitively they are given by kernels which are measures on \(\Pi_X\). The article is written in two levels: The first level of elementary differential geometry of bundles, connections and curvatures on \(C^\infty\)-manifolds, and the other that of formal geometry and ind-schemes, necessary to construct \(\widehat\Pi_X\). The issue of formal integration of Lie algebroids, and the groupoid analog of the fact that any Lie algebra over a field of characteristic \(0\) gives rise to a formal group, has been put to an appendix of the article. The author recalls main facts about Lie algebroids and their algebraic counterparts, Lie-Rinehart algebras. The main algebraic construction is that of a free Lie-Rinehart algebra generated by an anchored module. Finally the author gives an informal discussion of possible further directions motivated by this work. The article is mostly self contained and explains the theory in a reasonably easy and precise way. It is a very good introduction into a lot of areas in noncommutative geometry.