1: \begin{abstract}

2: We consider the problem of computing the Lebesgue volume of compact basic semi-algebraic sets. In full generality, it can be 

3: approximated as closely as desired by a converging hierarchy of upper bounds obtained by applying the 

4: Moment-SOS (sums of squares) methodology to a certain infinite-dimensional linear program (LP). At each step one solves a semidefinite relaxation of the LP which involves 

5: pseudo-moments up to a certain degree. Its dual computes a polynomial of same degree

6: which approximates from above the discontinuous indicator function of the set, hence with a typical Gibbs phenomenon

7: which results in a slow convergence of the associated numerical scheme. Drastic improvements have been observed by introducing in the initial LP

8: additional linear moment constraints obtained from a certain application of Stokes' theorem for integration

9: on the set. However and so far there was no rationale to explain this behavior. We provide a refined version of this 

10: extended LP formulation. When the set is the smooth super-level set of a single polynomial, we show that the

11: dual of this refined LP has an optimal solution which is a continuous function.

12: Therefore in this dual

13: one now approximates a continuous function

14: by a polynomial, hence with no Gibbs phenomenon, which explains and improves the already observed drastic acceleration of the convergence 

15: of the hierarchy. Interestingly, the technique of proof involves {recent} results on Poisson's partial differential equation (PDE).

16: \end{abstract}

17: