Stokes, Gibbs, and volume computation of semi-algebraic sets (Q2679606)

The authors of this paper concerns about the problem of computing the Lebesgue volume of compact basic semialgebraic sets, using the Moment-SOS methodology. This involves solving an infinite-dimensional linear program (LP) and obtaining the volume by taking the limit of a sequence of solutions as it converges. However, the convergence of the sequence can be slow due to the Gibbs phenomenon, which can impede the accuracy of the approximations. This issue can be resolved by introducing additional linear moment constraints obtained from an application of Stokes' theorem for integration on the set, which greatly improves convergence. While this approach has shown significant promise, the rationale behind its efficacy was unclear so far. The authors provide a refined version of the LP formulation, demonstrating that when the set is a smooth super-level set of a single polynomial, the dual of the refined LP has an optimal solution that is a continuous function. As a result, the dual approximates a continuous function with a polynomial, eliminating the Gibbs phenomenon and further accelerating convergence. The authors utilize recent results on Poisson's partial differential equation (PDE) in their proof of this technique. Overall, this paper presents a valuable contribution to the field of computational mathematics, providing a deeper understanding of the effective computation of Lebesgue volumes of semialgebraic sets.