Beyond Area: Perimeter Optimization in the Planar Skorokhod Embedding Problem
The planar Skorokhod embedding problem asks for a simply connected planar domain whose Brownian exit distribution has a prescribed real marginal $μ$. We investigate the minimization of conformal perimeter over the class of all such embeddings. We first establish the main existence theorem showing that whenever a finite-perimeter embedding exists, the infimum is attained (and vice versa). We derive necessary conditions for finiteness and sufficient conditions based on logarithmic integrability of the density. In particular, we demonstrate that finite area does not guarantee the existence of an embedding with finite perimeter. Although Gross's canonical embedding minimizes area, we construct an explicit family of domains for which Brownian symmetrization strictly increases perimeter, disproving its universal perimeter minimality. On the positive side, using Pruss's containment theorem and inequalities from convex geometry, we establish a general lower bound in terms of the convex hull of the Gross domain. Consequently, whenever the Gross embedding is bounded and convex, it is the unique perimeter minimizer. Finally, we formulate an open problem concerning the relationship between the existence of finite-perimeter embeddings and the regularity of Gross's canonical embedding.