nonabelian Stokes theorem

The *nonabelian Stokes theorem* is a generalization of the Stokes theorem for Lie algebra valued differential 1-forms and with integration of differential forms refined to parallel transport.

If $\hat A \in \Omega^1(D^2, \mathfrak{g})$ is a Lie algebra valued 1-form on the 2-disk then the parallel transport $\mathcal{P} \exp(\int_{S^1} A)$ of its restriction $A \in \Omega^1(S^1, \mathfrak{g})$ to the boundary circle, hence its holonomy (for a fixed choice of base point) is equal to a certain kind of adjusted 2-dimensional integral of its curvature 2-form $F_A$ over $D^2$.

In particular if $F_{\hat A} = 0$ then the holonomy of $A$ is trivial.

In terms of the notion of connection on a 2-bundle the nonabelian Stokes theorem says that if $A \in \Omega^1(X, \mathfrak{g})$ is a Lie algebra valued 1-form, then $(F_A, A)$ is a Lie 2-algebra valued 2-form with values in the inner derivation Lie 2-algebra $inn(\mathfrak{g})$ of $\mathfrak{g}$ whose curvature 3-form $H = \mathbf{d}_A F_A$ vanishes (which is the Bianchi identity for $F_A$) and its higher parallel transport exists. The 2-functorial source-target matching condition in this higher parallel transport is the statement of the nonabelian Stokes theorem.

For $F_A = 0$ the nonabelian Stokes theorem may be regarded as proving that the Lie integration of $\mathfrak{g}$ by “the path method” (see at Lie integration) is indeed the simply connected Lie group corresponding to $\mathfrak{g}$ by Lie theory.

For instance in

- Urs Schreiber, Konrad Waldorf,
*Smooth Functors vs. Differential Forms*(arXiv:0802.0663)

Revised on October 28, 2012 04:29:56
by Eric Forgy
(210.6.1.203)