Evolution of an extended Ricci flow system (Q1011996)

The author shows that Hamilton's Ricci flow and the static Einstein vacuum equations are closely connected by the following system of geometric evolution equations: \[ \begin{aligned} &\partial_tg=-2Rc(g)+2\alpha_n du\otimes du,\\ &\partial_tu=\Delta^gu,\end{aligned}\tag{*} \] where \(g(t)\) is a Riemannian metric, \(u(t)\) is a scalar function and \(\alpha_n\) a constant depending only on the dimension \(n\geq 3\). The author remarks that this provides an interesting and useful link from problems in low-dimensional topology and geometry to physical questions in general relativity. The author proves the existence of an entropy \(E\) such that the stationary points are solutions to the static Einstein vacuum equations and studies the above mentioned extended parabolic system, which is equivalent to the gradient flow of \(E\). For applications on noncompact asymptotically flat manifolds, the author proves short time existence for the system (*) on complete manifolds. For a close examination of the solutions, he provides a range a priori estimates. These include local interior estimates on balls for all derivatives of \(R\) and \(u\), but also global supremum bounds and time decay estimates. The author obtains a long time existence result from these estimates, stating that solutions continue to exist as long as the curvature \(R\) of \(g(t)\) stays bounded. From the variational structure the author obtains a monotonicity formula for the flow, which he uses to prove a noncollapsing result. This is crucial to obtain the injectivity radius bounds necessary for a compactness theorem. Using this result, compactness of a set of solutions satisfying natural bounds is obtained. In addition, the author applies the noncollapsing and the a priori estimates to prove that rapidly forming singularities of the flow at finite time can be rescaled to ancient solutions of the Ricci flow. This makes the study of singularity formation (and also the usage of surgery) possible. The author specifies that a main motivation to study (*) stems from its connection to general relativity. He says that an important issue in the numerical evolution of the Einstein equations is the construction of good initial data sets which have to satisfy the so-called constraint equations, and that in general, this is a hard problem (the author cites \textit{A. Rendall} [Theorems on existence and global dynamics for the Einstein equations, Living Rev. Relat. 8, No. 6 (2005) (URL) (cited on 12/12/05): \url{http://www.livingreviews.org/lrr-2005-6}]). The author says that a parabolic system could be used to improve these data sets, and since this should work in particular for static solutions, the system (*) is an interesting candidate for such a smoothing operator, and it should be possible to approximate static solutions by solutions to (*). The author mentions other applications.