Asymptotic evaluation of multidimensional oscillatory integrals

summary

The gist

Asymptotic evaluation of multidimensional oscillatory integrals develops an algebraic method to approximate these notoriously difficult integrals to any desired precision by constructing a…

In short

The paper develops an algebraic method to approximate difficult multidimensional oscillatory integrals. It constructs a finite-dimensional Gauss–Manin system that generates all necessary asymptotic information, including critical point series and Stokes constants. This allows for superasymptotic and hyperasymptotic approximations without needing gradient flows.

Key concepts

Gauss–Manin system
This is a finite-dimensional system of ordinary differential equations derived from the integrand. It is used algebraically to generate the asymptotic expansions of all critical points of the integral, providing a complete set of necessary information for approximation.
Stokes constants
These are integers determined by analyzing a Fuchsian system related to the Gauss-Manin system. They encode how adjacent saddle points interact and are crucial for determining which critical points actually contribute to the final integral sum.
Hyperasymptotics
This is a systematic method that improves approximations by resumming remainders along paths defined by the adjacency graph of saddle points, organized like Feynman diagrams. It allows for exponentially accurate approximations of oscillatory integrals.
Brieskorn reduction
This technique is used to reduce the integrand, particularly when dealing with polynomial or rational exponents. It transforms the problem into a linear system of ordinary differential equations in k that defines the Gauss-Manin system.

Terminology used across episodes

This episode discusses

The paper

Asymptotic evaluation of multidimensional oscillatory integrals · Read on arXiv

Samuel Crew, *, *Job Feldbrugge, ^3

QPerfect · Leung Center for Cosmology and Particle Astrophysics, National Taiwan University Institute for Advanced Study

Oscillatory integrals appear throughout physics, from diffraction and gravitational lensing to real-time path integrals. Their asymptotic evaluation is complicated by the need to determine which real and complex saddle points contribute, and how they are connected by Stokes phenomena. In multiple dimensions, this traditionally requires computing the steepest descent manifolds in a high-dimensional complex space. Here, we develop an algebraic alternative for integrals with polynomial and rational exponents. Using integration by parts and Brieskorn reduction, we construct a finite-dimensional Gauss-Manin system that generates the asymptotic expansions of all saddle points. Its Borel transform yields a Fuchsian system, whose monodromy determines the Stokes constants and hence the adjacency of the saddles. Combined with the recent South-East rule, this identifies the contributing saddles and enables superasymptotic and hyperasymptotic approximations without computing gradient flows. We demonstrate the method for the parabolic umbilic diffraction integral and a Lorentzian lens as examples, obtaining agreement to machine precision. The methods are implemented in an accompanying open-source Julia package IntegralResurgence.jl.

Transcript

Introduction to the show: ident: Mathematical Physics. The day in mathematical physics: rigorous models of physical systems.

Kai: We're looking at the paper: "Asymptotic evaluation of multidimensional oscillatory integrals".

Mira: Here's the gist of the paper. Asymptotic evaluation of multidimensional oscillatory integrals develops an algebraic method to approximate these notoriously difficult integrals to any desired precision by constructing a finite-dimensional Gauss–Manin system that determines all necessary asymptotic information Core Methodology The paper introduces an algebraic alternative for integrals with polynomial and rational exponents, using integration by parts and Brieskorn reduction to construct a finitedimensional Gauss–Manin system that generates the asymptotic expansions of all saddle points.

Kai: This system is transformed into a Fuchsian system in the Borel plane whose monodromy determines the Stokes constants, which are then used with the South-East rule to identify contributing saddles and enable superasymptotic and hyperasymptotic approximations without computing gradient flows.

Mira: The process involves reducing the integrand using Brieskorn reduction to satisfy a linear system of ordinary differential equations in k, known as the Gauss-Manin system.

Kai: Global Information Extraction The resurgent approximation requires three global pieces of information: (i) the asymptotic series of all critical points, (ii) the adjacency of the critical points encoded in Stokes constants, and (iii) which real and complex critical points actually contribute to the integral over R d.

Paper summary: Kai: So we're looking at this paper today, "Asymptotic evaluation of multidimensional oscillatory integrals." We just covered how they build this finite dimensional Gauss Manin system and the Borel plane transformation for finding Stokes constants.

Mira: Yeah, the main point is that they develop an algebraic alternative for integrals with polynomial and rational exponents to get asymptotic expansions to any desired precision.

Lev: They use integration by parts and Brieskorn reduction to construct this finite system that generates all the asymptotic expansions of all the saddle points, which is what makes it work.

Kai: That Gauss Manin system then gets transformed into a Fuchsian system in the Borel plane, and its monodromy gives us those Stokes constants which dictate the adjacency of the saddles.

Mira: They claim they extract three global pieces of information from this process: the asymptotic series for all critical points, and how they are connected by Stokes constants, and which real and complex critical points actually contribute to the integral over R d.

Lev: The method provides a simple recursion relation for the asymptotic series of all critical points, a Fuchsian system whose Borel transform yields those Stokes constants as integers with certified precision.

Kai: And they use that adjacency information with the South-East rule to pinpoint which critical points are relevant and enable superasymptotic and hyperasymptotic approximations without computing gradient flows.

Mira: They derive the asymptotic series for the moments of these integrals I j from this Gauss Manin equation, where the zeroth order in k reduces to an eigenequation MT j zero equals lambda j T j zero.

Lev: The higher order coefficients follow through Frobenius recursion, and they recover the transseries expansion of the periods I k by summing over all saddle points weighted by intersection numbers n j.

Kai: For superasymptotics, they truncate the series around each saddle point at its smallest term to get an exponentially accurate approximation with an error order e to the power of minus absolute value k F j l relative to the leading contribution of saddle point x j.

Mira: Then there's Stokes phenomena, where they map the integral to a Laplace integral in the Borel plane, and adjacent saddle points correspond to branch points that see each other on the Riemann surface.

Lev: They evaluate those Stokes constants K l j by comparing two keyhole paths or using a single loop formula derived from the Fuchsian system, which results in integer values.

Kai: This whole process lets them determine the adjacency graph and then use that to do hyperasymptotics by resumming remainders along paths organized as Feynman like diagrams.

Mira: The paper illustrates this with two examples: the parabolic umbilic diffraction integral, a genuinely two dimensional one with an exponent unbounded from below, and the Lorentzian lens model of wave optics.

Lev: In both cases, the method identifies the relevant complex saddle points and their adjacency relations to reproduce the integral to machine precision.

Conclusion: Kai: So wrapping up this paper "Asymptotic evaluation of multidimensional oscillatory integrals," it seems like they've put together a really powerful algebraic framework.

Mira: They connect the Brieskorn lattice with the Birkhoff problem and the resurgence theory, showing how Stokes constants are topological invariants related to intersection numbers of vanishing cycles.

Lev: The adjacency graph itself describes direct Stokes couplings, which is what lets them figure out which saddle points interact and contribute to the integral.

Kai: So basically, they've given us a way to get an exponentially accurate approximation without ever computing a single steepest descent manifold in high dimensional complex space.

Mira: They successfully identify relevant complex saddle points and their adjacency algorithmically, which fulfills the vision of hyperasymptotics for multidimensional integrals.

Lev: This approach provides a computable description of particle interactions where saddle point expansions behave as weakly interacting sectors coupled through Stokes phenomena.

Kai: It's a way to get high precision results for integrals in regimes where standard numerical methods just can't handle the complexity anymore.

More episodes

← Home