Symbolic Reduction of Multi-loop Feynman Integrals via Generating Functions
Bo Feng a,b,c,d, Xiang Li e, Yuanche Liu g, Yan-Qing Ma e,f, Yang Zhang g,d,f · 2025
Read Symbolic Reduction of Multi-loop Feynman Integrals via Generating Functions by Bo Feng a,b,c,d, Xiang Li e, Yuanche Liu g, Yan-Qing Ma e,f, Yang Zhang g,d,f (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-2509-21769.txt.
Introduction
Bo Feng and four co-authors present a new way to simplify Feynman integrals in this 2025 paper. Such integrals sit at the heart of particle physics calculations, and the multi-loop ones are notoriously hard. The standard approach builds huge systems of relations that grow explosively with the size of the problem. Their method aims to cut every integral down to a small set of master integrals.
The key tool is a generating function, a single object that packages many integrals at once. Equations for that function yield step-by-step rules for reducing the integrals one by one. The team turns those rules into an efficient algorithm. It avoids the blow-up of the older relation-based approach. That makes it possible to reduce high-rank, multi-loop integrals that modern precision calculations in quantum field theory depend on.
Catalog record
| CREATOR | Bo Feng a,b,c,d, Xiang Li e, Yuanche Liu g, Yan-Qing Ma e,f, Yang Zhang g,d,f |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 86 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-21769
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Symbolic Reduction of Multi-loop Feynman Integrals via Generating Functions Bo Feng a,b,c,d, Xiang Li e, Yuanche Liu g, Yan-Qing Ma e,f, Yang Zhang g,d,f Abstract We introduce a novel, systematic method for the complete symbolic reduction of multi-loop Feynman integrals, leveraging the power of generating functions. The differential equations governing these generating functions naturally yield symbolic recurrence relations. We develop an efficient algorithm that utilizes these recurrences to reduce integrals to a minimal set of master integrals. This approach circumvents the exponential growth of traditional Integration-by-Parts relations, enabling the reduction of high-rank, multi-loop integrals critical for state-of-the-art calculations in perturbative quantum field theory. USTC-ICTS/PCFT-25-37 I Introduction Loop-level Feynman diagrams play a central role in perturbative quantum field theories (QFTs). In the era of precision physics, The current era of precision physics presents an urgent demand for the calculation of multi-loop amplitudes, which are crucial for collider phenomenology, gravitational wave predictions from black hole scattering, as well as formal aspects of QFT. In multi-loop computations, a key step is to reduce all Feynman integrals to a finite, linearly independent set consisting of master integrals. This step, known as Feynman integral reduction, is frequently the computationally intensive bottleneck of perturbative QFT calculation. Conventionally, integral reduction is carried out using the integration-by-parts (IBP) identities Tkachov 1981 ; Chetyrkin and Tkachov 1981 via the linear reduction algorithm named after Laporta Laporta 2000 . However, for most cutting-edge multi-loop computations, the IBP method inevitably involves a vast linear system, which is prohibitively difficult to solve via linear reduction. There are several public software packages for Feynman integral reduction, implementing the Laporta algorithm along with various modern techniques: AIR Anastasiou and Lazopoulos 2004 , FIRE Smirnov 2008 ; Smirnov and Smirnov 2013 ; Smirnov 2015 ; Smirnov and Chukharev 2020 , LiteRed Lee 2012 ; Lee 2014 , Reduze Studerus 2010 ; von Manteuffel and Studerus 2012 , FiniteFlow Peraro 2016 , Kira Maierhöfer et al. 2018 ; Maierhöfer and Usovitsch 2018 ; Maierhöfer and Usovitsch 2020 ; Klappert et al. 2021a ; Lange et al. 2025 , FireFly Klappert and Lange 2020 ; Klappert et al. 2021b , RaTracer Magerya 2022 , Blade Guan et al. 2020 ; Liu and Ma 2022 ; Guan et al. 2025 , and NeatIBP Wu et al. 2024 ; Wu et al. 2025 . Notable advances that have boosted linear reduction include the finite-field method von Manteuffel and Schabinger 2015 , the application of computational algebraic geometry Gluza et al. 2011 ; Larsen and Zhang 2016 ; Wu et al. 2024 ; Wu et al. 2025 , and the block-triangular form Guan et al. 2020 ; Liu and Ma 2022 ; Guan et al. 2025 . An alternative framework is provided by intersection theory Mastrolia and Mizera 2019 ; Chestnov et al. 2022 ; Brunello et al. 2025 , formulating Feynman integral reduction in terms of twisted cohomology. Despite these developments, the capabilities of existing reduction methods are still insufficient to meet the demands of upcoming precision experiments, such as the High-Luminosity Large Hadron Collider (HL-LHC), the Laser Interferometer Space Antenna (LISA), the Taiji and Tianqin projects. The main bottleneck of integration-by-parts (IBP) reduction arises in multi-loop integrals with high-degree numerators or high-power propagators, where the number of IBP relations grows exponentially. This issue becomes particularly severe in recent cutting-edge multi-loop computations, such as the 5 pm 1 SF black-hole scattering Driesse et al. 2025 and the three-loop five-point Feynman integrals Liu et al. 2025 . Tackling reduction problems at the three- and four-loop orders, or for two-loop multi-leg processes, necessitates new conceptual breakthroughs. Recurrence rules Smirnov and Smirnov 2006a ; Lee 2012 provide an efficient alternative to IBP reductions, particularly for the challenging task of reducing multi-loop integrals with high-rank numerators or elevated propagator powers. The implements of recurrence rules include methods based on gröbner basis Tarasov 1998 ; Gerdt 2004 ; Smirnov and Smirnov 2006a ; Smirnov and Smirnov 2006b ; Smirnov 2006 and heuristic algorithms Lee 2012 ; Lee 2014 . However, a significant drawback was the computational cost associated with the recurrence rule derivation, which has limited their widespread adoption. In this work, we present a novel and highly efficient method, based on generating functions Feng 2023 ; Guan et al. 2023 , to systematically derive complete symbolic recurrence relations for the reduction of Feynman integrals. By expressing Feynman integrals in terms of generating functions, integration-by-parts (IBP) relations are translated into differential equations for these generating functions. The reduction of Feynman integrals is then transformed into the task of reducing a differential operator with respect to these differential equations. We introduce a new approach for differential operator reduction, based on operator grading and integrability conditions, which allows us to efficiently tobtain symbolic reduction rules for arbitrary Feynman integrals. This paper is organized as follows. We first review the definition of generating functions for Feynman integral reduction. We then focus on our new systematic method for obtaining symbolic reduction rules. As illustrations, we apply this method to three multi-loop reduction examples: the sunset diagram, and planar and non-planar double-box diagrams. During the preparation of this manuscript, we became aware of the recent paper Smith and Zeng 2025 , which presents an efficient method for generating Feynman integral reduction rules using syzygies. In this work, we emphasize that our method uses only fundamental IBP relations and does not rely on syzygies. II Generating Functions for Feynman integrals For L -loop integrals with E independent external momenta, a family of scalar FIs can be represented as \displaystyle I(\vec{\nu})=\int\prod_{i=1}^{L}\frac{\mathrm{d}^{D}\ell_{i}}{\mathrm{i}\pi^{D/2}}\frac{\mathcal{D}_{K+1}^{-\nu_{K+1}}\cdots\mathcal{D}_{N}^{-\nu_{N}}}{\mathcal{D}_{1}^{\nu_{1}}\cdots\mathcal{D}_{K}^{\nu_{K}}},\penalty\ \penalty\ \penalty\ (1) where D=4-2\epsilon represents the spacetime dimension, \ell_{i} are the loop momenta, \mathcal{D}_{1},\ldots,\mathcal{D}_{K} are the inverse propagators, \mathcal{D}_{K+1},\ldots,\mathcal{D}_{N} are the irreducible scalar products (ISPs) introduced to form a complete basis and N={L(L+1)\over 2}+LE . In ( 1 ), \nu_{1},\ldots,\nu_{K} can be arbitrary integers (i.e, positive, negative or zero), while \nu_{K+1},\ldots,\nu_{N} can only be non-positive integers. An integral is said to belong to the top-sector of this family if all its propagator indices are positive, i.e., \nu_{1},\ldots,\nu_{K}>0 . To organize all integrals within a sector, we introduce a generating function defined as: \displaystyle G_{\vec{\mu}}(\eta)=\int\prod_{i=1}^{L}\frac{\mathrm{d}^{D}\ell_{i}}{\mathrm{i}\pi^{D/2}}{e^{\sum_{j=1}^{N}(1-\mu_{j})\eta_{j}s_{0}^{-1}\mathcal{D}_{j}}\over\prod_{i=1}^{N}(\mathcal{D}_{i}-s_{0}\eta_{i})^{\mu_{i}}},\penalty\ \penalty\ \penalty\ (2) where \vec{\mu}=(\mu_{1},...,\mu_{N}) consists of \mu_{i}=0,1 and s_{0} is an auxiliary scale that make \eta_{i} dimensionless. For top-sector, we set \mu_{1},\cdots,\mu_{K}=1 and \mu_{K+1},\cdots,\mu_{N}=0 and G_{\vec{\mu}} can be expanded as: \displaystyle G_{\vec{\mu}}(\vec{\eta}) \displaystyle= \displaystyle\sum_{\vec{n}\geq 0}\vec{\eta}^{\vec{n}}F_{\vec{n}},\penalty\ \penalty\ \penalty\ (3) with \displaystyle F_{\vec{n}} \displaystyle= \displaystyle\left(\prod_{j=K+1}^{N}{1\over n_{j}!}\right)s_{0}^{\sum_{i=1}^{K}n_{i}-\sum_{j=K+1}^{N}n_{j}} (4) \displaystyle\times \displaystyle I(...,n_{K}+1,-n_{K+1},...,-n_{N})\penalty\ . The formalism ( 3 ) has mapped Feynman integrals to the expansion coefficients of generating function, labeled by the lattice points \vec{n} . We define the degree of F as \sum_{i=1}^{N}n_{i} . Consequently, the reduction problem of Feynman integrals is transformed into finding relations among these coefficients, which can be derived from differential equations(DEs) satisfied by the generating function. Since the generating function encodes all Feynman integrals in the family, it is sufficient to construct IBP relations for the generating function alone, which we call as the fundamental IBP relations. These relations yield DEs of the form: \displaystyle\sum_{t}c_{t}\hat{O}_{t}G_{\vec{\mu}}(\vec{\eta})=B\penalty\ ,\penalty\ \penalty\ \penalty\ \hat{O}_{t}=\prod_{i=1}^{N}\eta_{i}^{(\vec{a}_{t})_{i}}{\partial^{(\vec{b}_{t})_{i}}\over\partial\eta_{i}^{(\vec{b}_{t})_{i}}\penalty\ },\penalty\ (5) where B is the function of generating functions of subsectors and c_{t} are coefficients. We define the degree of an operator as \sum_{i=1}^{N}(\vec{o})_{i} , where the index vector is \vec{o}\equiv\vec{b}-\vec{a} .