Probabilistic representation and uniqueness results for measure-valued solutions of transport equations (Q996856)

This article is concerned with the Cauchy problem for a multidimensional linear non-homogeneous transport equation in divergence form of type \[ \partial_t \mu_t + D_x \cdot (v_t \mu_t) = c_t \mu_t \qquad \text{ in }\quad (0,T)\times \mathbb{R}^d. \] Based on the method of characteristics an explicit and an implicit representation formula for the unique solution \(\{\mu_t\}\) in the Schwartz distribution sense are developed in the case of a regular vector field \(v\). Via regularization, these formulas and some tightness arguments are applied further on to obtain a very general probabilistic representation for measure valued solutions in the case when the initial datum is a bounded measure and the coefficient vector field \(v\) is no more regular, but satisfies suitable summability assumptions w.r.t. the solution. In the last part, the author connects the concept of renormalized solutions and probabilistic solutions (in the distributional sense) of the equation. In the sequel uniqueness results for solutions of the initial value problem can be derived from the uniqueness of the characteristic curves associated to \(v\) through the theory of the probabilistic representation previously exposed.