The Theory of Causal Fermion Systems

Flows of Measures

Flows of Measures

The causal action principle or, more generally, causal variational principles give rise to flows of measures. These flows can be constructed even in the non-smooth setting using the method of minimizing movements. So far, these techniques have been used in two different contexts:

Action-driven flows

With the goal of constructing minimizers of the causal variational principles, one wants to construct a flow of measures such that the action decreases along the flow lines. In the smooth setting, this is similar to a gradient flow, where the measure flows in the “direction of steepest descent.” In the general non-smooth situation, the flow is constructed in [FG25] using the method of minimizing movements. This method makes it possible to construct approximate solutions of the Euler-Lagrange equations even in the infinite-dimensional case.

The isoperimetric flow

Flows of measures can also be used in order to study geometric properties of a given space or spacetime $M$. Suppose that $M$ is described by a measure $\mu$ on a smooth manifold $\F$. Then the flow is described by a family of measures $(\rho_t)_{t \in \R}$ with $0 \leq \rho(t) \leq \mu$. In [FGK26] the isoperimetric flow is introduced as the gradient flow corresponding to the area functional $A(\rho)$ defined by
$$ A(\rho) = \int_M d\rho(x) \int_M d(\mu – \rho)(y)\: \L(x,y) \:. $$
This flow is constructed using minimizing movements as a Lipschitz curve of measures with prescribed volume change. The resulting flow makes it possible to introduce notions of local dimension and synthetic scalar curvature.