Schwinger-Dyson approximants
Bartomeu Fiol, Elena Gijón, and Unai Lejarza Alonso · 2025
Read Schwinger-Dyson approximants by Bartomeu Fiol, Elena Gijón, and Unai Lejarza Alonso (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-05665.txt.
Introduction
Bartomeu Fiol, Elena Gijon, and Unai Lejarza Alonso revisit the Schwinger-Dyson equations in the simplest possible setting: a zero-dimensional field theory. These equations link all of a theory's correlation functions, but they form an endless tower, so users truncate them. The authors ask what solves the truncated system exactly.
They argue the solutions are rational approximants, built by strikingly simple recursion rules that never use the usual perturbative expansion. For the lowest correlations these match the classic Padé approximants, but for higher ones they differ, a split the authors tie to which functions are of Stieltjes type. As the truncation grows, the approximants provably converge to the full non-perturbative answers.
Catalog record
| CREATOR | Bartomeu Fiol, Elena Gijón, and Unai Lejarza Alonso |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 67 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-05665
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Schwinger-Dyson approximants Bartomeu Fiol, Elena Gijón, and Unai Lejarza Alonso Abstract We revisit the solution to the Schwinger-Dyson equations in the simple case of the 0-dimensional 1 2 m 2 ϕ 2 + λ 4 ϕ 4 \frac{1}{2}m^{2}\phi^{2}+\frac{\lambda}{4}\phi^{4} theory with m 2 > 0 m^{2}>0 and λ ≥ 0 \lambda\geq 0 . We argue that the truncated Schwinger-Dyson equations are solved by rational approximants to all n-point functions ⟨ ϕ 2 k ⟩ \langle\phi^{2k}\rangle , and provide strikingly simple recursive relations for them. These rational approximants are constructed without any reference to ordinary perturbative expansions. They turn out to be Padé approximants for ⟨ ϕ 2 ⟩ \langle\phi^{2}\rangle and for half of the truncations in the case of ⟨ ϕ 4 ⟩ \langle\phi^{4}\rangle , but they are not Padé approximants for higher n-point functions. This difference is related to the fact that ⟨ ϕ 2 ⟩ \langle\phi^{2}\rangle and ⟨ ϕ 4 ⟩ \langle\phi^{4}\rangle are Stieltjes functions, while higher n-point functions are not. We prove that as the size of the truncation tends to infinity, these rational approximants converge to the full non-perturbative n-point functions for all positive values of the coupling λ \lambda . Thus, in the example studied in this work, these new rational approximants are much easier to derive than the usual Padé approximants, and when different, they are better suited to approximate the full non-perturbative n-point functions. 1 Introduction There is more to Quantum Field Theory (QFT) than perturbative expansions. It has been known since the early days of QFT that generic perturbative expansions are asymptotic Dyson:1952tj ; LeGuillou:1990nq , so they may miss qualitatively important features of the theory. This has motivated the development of a variety of non-perturbative approaches to study QFT. Among those approaches, one of the earliest ones are the Schwinger-Dyson (SD) equations Dyson:1949ha ; Schwinger:1951ex ; Schwinger:1951hq : these are an infinite set of equations relating the n-point functions of the theory. For physically relevant four-dimensional gauge QFTs like QED or QCD, the full set of SD equations is impossibly hard to solve. One then resorts to various truncation schemes, and has to deal with issues related to gauge-fixing and renormalization, see Mandelstam:1979xd ; Cornwall:1981zr ; Cornwall:1989gv ; Curtis:1990zs ; Roberts:1994dr ; Alkofer:2000wg for a necessarily incomplete list of references. To explore truncation schemes for the Schwinger-Dyson equations, a possible strategy is to test different approaches on simpler, lower dimensional QFTs where exact results are known, see e.g. Bender:1976xh ; Dorey:1989mr ; Adam:1994by . Taking this philosophy to the extreme, there has been plenty of work applying the Schwinger-Dyson equations to 0-dimensional theories, where path integrals reduce to ordinary integrals Bender:1988bp ; Okopinska:1990pt ; Garcia:1996np ; Bender:2022eze ; Bender:2023ttu ; Peng:2024azv ; Banks:2024ydh . While in these toy models the thorny issues of gauge-fixing or renormalization are simply absent, their solutions already display non-analytic behavior, which makes them an interesting training ground. In this work, we join this line of research and revisit the familiar case of the zero-dimensional Euclidean \lambda^{4} theory, S=\frac{1}{2}m^{2}\phi^{2}+\frac{\lambda}{4}\phi^{4} (1) in the case m^{2}>0 and \lambda\geq 0 . The aim of this work is to use the Schwinger-Dyson equations to obtain approximants to the exact n-point functions of the theory for the choice of integration contour along the real axis, \langle\phi^{2k}\rangle=\frac{\int_{-\infty}^{\infty}d\phi\,\,\phi^{2k}\,\,e^{-S(\phi)}}{\int_{-\infty}^{\infty}d\phi\,\,e^{-S(\phi)}} (2) and to demonstrate the convergence of these new approximants to the exact non-perturbative n-point functions, for all positive values of the coupling \lambda . We will argue that for the theory at hand, the truncation of the infinite tower of SD equations to the first n SD equations is solved by rational approximants to the 2k-point functions, that we call the Schwinger Dyson approximants, and denote by \langle\phi^{2k}\rangle^{n}_{\text{SD}} . These rational approximants are not entirely new: to the best of our knowledge, they first appeared in Bender:1988bp , where it was claimed that their validity is limited to the weak coupling region, a claim we disagree with. Much more recently, these rational approximants appeared again in Konosu:2024zrq , who arrived at them from a superficially different starting point: the homotopy algebra approach to QFT Hohm:2017pnh ; Jurco:2018sby . The arguments in Bender:1988bp and the present work show that the main result of Konosu:2024zrq can be derived immediately from the SD equations, without having to rely on the homotopy algebra framework 1 1 1 See Okawa:2022sjf ; Konosu:2023psc ; Konosu:2024dpo for proofs that the correlation functions obtained with the homotopy algebra approach to QFT satisfy the SD equations. . On the other hand, the authors of Konosu:2024zrq claim but do not prove that these rational approximants capture also the strong coupling regime of the exact n-point functions. Making precise this claim was the original motivation for the present work, and our two main results are an explicit recursive definition of the novel rational approximants, and the proof of their convergence to the exact n-point functions, for all positive values of the coupling. The structure of the paper is as follows. In section 2 we present the exact n-point functions \langle\phi^{2k}\rangle and recall the well-known power series, \langle\phi^{2k}\rangle_{\text{weak}} and \langle\phi^{2k}\rangle_{\text{strong}} . We then discuss the analytic properties of the exact \langle\phi^{2k}\rangle and point out a difference that will play a major role in the rest of the paper: as functions of \lambda in the complex plane \langle\phi^{2}\rangle(\lambda) , \langle\phi^{4}\rangle(\lambda) are Stieltjes functions 2 2 2 The definition and basic facts about Stieltjes functions are reviewed in appendix A . , while higher n-point functions are not Stieltjes. Nevertheless, we show that \frac{\langle\phi^{2k+2}\rangle}{\langle\phi^{2k}\rangle} are Stieltjes functions for k\geq 0 . This has the important implication that while \langle\phi^{6}\rangle are higher n-point functions are not Stieltjes, they are equal to the product of a finite number of Stieltjes functions, \langle\phi^{2k}\rangle=\frac{\langle\phi^{2k}\rangle}{\langle\phi^{2k-2}\rangle}\frac{\langle\phi^{2k-2}\rangle}{\langle\phi^{2k-4}\rangle}\dots\frac{\langle\phi^{4}\rangle}{\langle\phi^{2}\rangle}{\langle\phi^{2}\rangle}. (3) This observation will provide a useful perspective on the rational approximants we discuss next. We then turn to the main topic of this work, and solve the truncated Schwinger-Dyson equations of the theory ( 1 ) by rational approximants defined recursively. Namely, introducing the families of recursive polynomials in \lambda labelled by k=0,1,2,\dots , p_{n}^{(k)}(\lambda)=\begin{cases}0&n<k+1\\ 1&n=k+1\\ p_{n-1}^{(k)}(\lambda)+(2n-3)\lambda p_{n-2}^{(k)}&n>k+1\end{cases} (4) we define the rational approximants \langle\phi^{2k}\rangle_{\text{SD}}^{n}=(2k-1)!!\,\frac{p_{n}^{(k)}(\lambda)}{p_{n}^{(0)}(\lambda)} (5) which satisfy the first n-1 Schwinger-Dyson equations. As discussed in the main text, these rational approximants have a number of appealing properties, e.g. for a given truncation, approximants to different 2k-point functions all have the same denominator, and they don’t present spurious zeros or poles on the positive real axis. In section 3 we proceed to compare the new rational approximants \langle\phi^{2k}\rangle_{\text{SD}}^{n} to the most familar rational approximants, the Padé approximants obtained from asymptotic weak coupling series \langle\phi^{2k}\rangle_{\text{weak}} . We rigorously prove that for the 2-point function, \langle\phi^{2}\rangle_{\text{SD}}^{n} are precisely the familiar [N,N] and [N,N+1] Padé approximants. For the 4-point function, half of the new approximants \langle\phi^{4}\rangle_{\text{SD}}^{n} are precisely the [N,N+1] Padé approximants, while the other half don’t coincide with any Padé approximant. Notably, starting with \langle\phi^{6}\rangle_{\text{SD}}^{n} , the new approximants do not coincide with any Padé approximants to \langle\phi^{2k}\rangle_{\text{weak}} . Rather, they are products of Padé approximants to the Sieltjes functions in ( 3 ). This provides a compelling perspective on the usefulness of the SD approximants: if a function is Stieltjes, Padé approximants are the most convenient