Papers
Topics
Authors
Recent
Search
2000 character limit reached

Extended Partition Function Overview

Updated 10 July 2026
  • Extended Partition Function is a concept that enlarges the traditional partition function with additional variables, couplings, and geometric constraints to capture new observables.
  • It unifies diverse areas such as holography, Euclidean gravity, integrable hierarchies, and combinatorics by systematically refining classical models and introducing new counting rules.
  • This framework facilitates novel computational and analytical techniques, including matrix model refinements, Pfaffian representations, and continued fraction methods, to explore complex physical and mathematical systems.

“Extended partition function” is a field-dependent term for constructions that enlarge an ordinary partition function by adjoining extra variables, extra couplings, geometric constraints, refined symmetry data, or new counting rules. In the literature considered here, the extension may identify holographic complexity with a grand partition function, impose a fixed spatial-volume constraint in Euclidean gravity, refine open intersection-theoretic and KP tau-functions by extra times, or generalize Euler’s partition function to rectangle tilings, multiplicative statistics on partitions, and largest-part constrained families (Sun et al., 2019, Wu et al., 13 Dec 2025, Wang, 21 Nov 2025, Bertola et al., 2014, Gajdzica et al., 24 Sep 2025, Bal et al., 2021). This suggests that the phrase does not denote a single canonical object; rather, it denotes a recurrent strategy for extending a baseline partition-theoretic framework.

1. Terminological range and recurrent extension mechanisms

Several mathematically distinct constructions appear under the same label. They can be organized by the object being extended and by the mechanism of extension.

Domain Extension mechanism Representative result
Holography and thermodynamics Grand canonical variables and complexity growth C˙=TlnZ\dot{\mathcal C}=T\ln\mathcal Z (Sun et al., 2019)
Euclidean gravity Fixed-volume constraint and constrained instantons IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G) (Wu et al., 13 Dec 2025)
Open intersection theory Extra couplings sis_i and matrix refinement ZNo,ext=ZN\mathcal Z_N^{o,\mathrm{ext}}=Z_N under Miwa parametrization (Wang, 21 Nov 2025)
rr-reduced KP hierarchy Extra KP times and generalized string equation ZE(tKP)=Z(t)Y(tKP)Z_E(t^{KP})=Z(t)\,\mathcal Y(t^{KP}) (Bertola et al., 2014)
Torus-knot/Hurwitz tau-functions Fractional-index times and topological recursion EO recursion reproduces the nn-point functions of ZextZ^{\mathrm{ext}} (Dunin-Barkowski et al., 2020)
Enumerative combinatorics Geometric, multiplicative, or constrained partition rules p(m,n)p(m,n), p(λ)p(\lambda), and IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)0 (Gajdzica et al., 24 Sep 2025, Bessenrodt et al., 2014, Johnson et al., 11 Oct 2025)
Numerical statistical mechanics Temperature promoted to a sampled variable Single-run nested sampling for temperature-dependent potentials (Maillard et al., 2 Sep 2025)

A common pattern is the replacement of a single generating object by a richer one: a partition function depending on more variables, a partition function evaluated in a larger ensemble, or a counting function defined on a larger class of combinatorial configurations. A second pattern is representational: several papers extend not the underlying physical theory alone but the formalism used to compute its partition function, as in matrix-model identifications, Pfaffian encodings, Berezin integrals, continued fractions, and matrix-recursive calculi (Sinclair, 2011, Gobron, 2013, Campbell, 2023).

2. Thermodynamic and gravitational formulations

In holography, the paper “Complexity growth rate, grand potential and partition function” formulates an explicit “Complexity/Grand potential/Partition Function” framework, or CV 3.0, by combining the late-time CV 2.0 relation IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)1 with the grand-canonical identities IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)2 and IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)3. The resulting ansatz is

IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)4

and, after setting IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)5,

IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)6

The framework is then applied to original SYK, complex SYK, higher-dimensional coupled SYK islands, SYK wormholes, Schwarzschild–AdS, and RN–AdS geometries. On the bulk side, the semiclassical relation IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)7 gives IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)8, so complexity growth is directly linked to the Euclidean action. The same framework is used to discuss Hawking–Page behavior, first-order small/large black-hole transitions, sign/stability issues, Lloyd-type bounds, and the “second law of complexity” through IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)9 (Sun et al., 2019).

A different gravitational extension appears in “On the gravitational partition function under volume constraints.” There the Euclidean action is augmented by a Lagrange multiplier imposing a fixed spatial volume on each Euclidean time slice,

sis_i0

The resulting fixed-volume partition function

sis_i1

is evaluated semiclassically over a family of volume-constrained Euclidean geometries. The paper distinguishes a massless VCEG, a massive VCEG with an artificial boundary at sis_i2, and an extended VCEG in which that boundary is replaced by a throat and the geometry acquires two horizons. For the extended VCEG, the on-shell action becomes

sis_i3

with conical defects at the two horizons except at a critical mass sis_i4 where sis_i5 and the geometry is smooth. These objects are interpreted as constrained gravitational instantons, and their saddle contributions produce an analogy with Euclidean Schwarzschild–de Sitter in which the fixed-volume constraint plays a role akin to a cosmological constant (Wu et al., 13 Dec 2025).

3. Integrable hierarchies, matrix models, and refined tau-functions

In open intersection theory, the paper “The identification of the extended refined open partition function and the Kontsevich-Penner matrix model” introduces the matrix model sis_i6 by replacing the scalar complex integral in the Buryak–Tessler open partition function with a complex matrix integral over sis_i7 and by inserting couplings sis_i8 through

sis_i9

Under the Miwa parametrization

ZNo,ext=ZN\mathcal Z_N^{o,\mathrm{ext}}=Z_N0

the model reduces exactly to the Kontsevich–Penner matrix model ZNo,ext=ZN\mathcal Z_N^{o,\mathrm{ext}}=Z_N1, establishing

ZNo,ext=ZN\mathcal Z_N^{o,\mathrm{ext}}=Z_N2

for all ZNo,ext=ZN\mathcal Z_N^{o,\mathrm{ext}}=Z_N3. The paper derives this by transforming the original open partition function with the Harish–Chandra–Itzykson–Zuber formula and operational calculus, and it relates the refined open model to Virasoro constraints and the KP/open-KdV hierarchy (Wang, 21 Nov 2025).

The paper “The partition function of the extended ZNo,ext=ZN\mathcal Z_N^{o,\mathrm{ext}}=Z_N4-reduced Kadomtsev-Petviashvili hierarchy” extends the closed ZNo,ext=ZN\mathcal Z_N^{o,\mathrm{ext}}=Z_N5-spin partition function ZNo,ext=ZN\mathcal Z_N^{o,\mathrm{ext}}=Z_N6 to a solution of the extended ZNo,ext=ZN\mathcal Z_N^{o,\mathrm{ext}}=Z_N7-reduced KP hierarchy by adjoining the missing KP times and imposing a generalized string equation. The extended partition function is written

ZNo,ext=ZN\mathcal Z_N^{o,\mathrm{ext}}=Z_N8

and satisfies

ZNo,ext=ZN\mathcal Z_N^{o,\mathrm{ext}}=Z_N9

The factor rr0 is constructed from the wave function by the contour integral

rr1

where rr2 is the asymptotic series of a Pearcey-type integral solution. The construction generalizes Buryak’s rr3 extended open KdV solution to arbitrary rr4 and is specified uniquely, up to multiplicative constant, by the generalized string equation and normalization (Bertola et al., 2014).

A third integrable extension is the extended Ooguri–Vafa partition function for colored HOMFLY–PT polynomials of torus knots. It is defined by a fermionic vacuum expectation value with ordinary times rr5 and fractional-index times rr6,

rr7

After a change of variables, it becomes a hypergeometric KP tau-function, and EO topological recursion on the torus-knot spectral curve reproduces the correlation differentials rr8 extracted from rr9 (Dunin-Barkowski et al., 2020).

An adjacent algebraic development is the SHZE(tKP)=Z(t)Y(tKP)Z_E(t^{KP})=Z(t)\,\mathcal Y(t^{KP})0-based analysis of Nekrasov instanton partition functions for ZE(tKP)=Z(t)Y(tKP)Z_E(t^{KP})=Z(t)\,\mathcal Y(t^{KP})1 ZE(tKP)=Z(t)Y(tKP)Z_E(t^{KP})=Z(t)\,\mathcal Y(t^{KP})2 linear quivers. There the extension lies in the symmetry algebra rather than in a new partition-function variable: SHZE(tKP)=Z(t)Y(tKP)Z_E(t^{KP})=Z(t)\,\mathcal Y(t^{KP})3 is a deformation of ZE(tKP)=Z(t)Y(tKP)Z_E(t^{KP})=Z(t)\,\mathcal Y(t^{KP})4 containing ZE(tKP)=Z(t)Y(tKP)Z_E(t^{KP})=Z(t)\,\mathcal Y(t^{KP})5, and its generators act by adding or removing boxes in Young-diagram data, yielding infinite recursion relations for the instanton partition function and reproducing Heisenberg and Virasoro constraints compatible with the AGT correspondence (Kanno et al., 2013).

4. Enumerative and number-theoretic generalizations

In combinatorics, “Rectangle partitions generalizing integer partitions” defines ZE(tKP)=Z(t)Y(tKP)Z_E(t^{KP})=Z(t)\,\mathcal Y(t^{KP})6 as the number of partitions of the ZE(tKP)=Z(t)Y(tKP)Z_E(t^{KP})=Z(t)\,\mathcal Y(t^{KP})7 rectangle into axis-aligned rectangular blocks with positive integer side lengths, counted up to equality of multisets of block sizes and with ZE(tKP)=Z(t)Y(tKP)Z_E(t^{KP})=Z(t)\,\mathcal Y(t^{KP})8 identified with ZE(tKP)=Z(t)Y(tKP)Z_E(t^{KP})=Z(t)\,\mathcal Y(t^{KP})9. The one-dimensional specialization recovers Euler’s partition function: nn0 The paper develops bounds and asymptotics for nn1, including

nn2

restricted counts nn3 with explicit recurrences and quasi-polynomial behavior, nn4-ary rectangular partitions with congruence properties, and symmetric variants such as nn5 with

nn6

Here the extension is geometric: parts become integer-sided rectangles, while exact tilability replaces mere summation of integers (Gajdzica et al., 24 Sep 2025).

A more algebraic extension appears in the Partition-Frequency Enumeration matrix. The paper “The Partition-Frequency Enumeration Matrix” defines an infinite upper-triangular number-theoretic matrix whose finite truncations nn7 couple partition-like coefficients nn8 and part-frequency statistics nn9 through

ZextZ^{\mathrm{ext}}0

and

ZextZ^{\mathrm{ext}}1

For ordinary partitions, ZextZ^{\mathrm{ext}}2 if ZextZ^{\mathrm{ext}}3 and ZextZ^{\mathrm{ext}}4 otherwise, yielding

ZextZ^{\mathrm{ext}}5

The same calculus extends to arbitrary generating functions, Weierstrass products, and recurrences for ZextZ^{\mathrm{ext}}6, ZextZ^{\mathrm{ext}}7, ZextZ^{\mathrm{ext}}8, ZextZ^{\mathrm{ext}}9, and p(m,n)p(m,n)0, and it embeds Ramanujan-type congruences into infinite families (Bal et al., 2021).

A third line of work extends the partition function directly to functions on partitions. In “Maximal multiplicative properties of partitions,” the extension is

p(m,n)p(m,n)1

for p(m,n)p(m,n)2. The paper proves that for p(m,n)p(m,n)3, p(m,n)p(m,n)4, the unique maximizer of p(m,n)p(m,n)5 is determined by p(m,n)p(m,n)6: all p(m,n)p(m,n)7’s when p(m,n)p(m,n)8, one p(m,n)p(m,n)9 and the rest p(λ)p(\lambda)0’s when p(λ)p(\lambda)1, one p(λ)p(\lambda)2 and the rest p(λ)p(\lambda)3’s when p(λ)p(\lambda)4, and one p(λ)p(\lambda)5, one p(λ)p(\lambda)6, and the rest p(λ)p(\lambda)7’s when p(λ)p(\lambda)8 (Bessenrodt et al., 2014).

Largest-part constrained variants furnish another explicit extension. “On recent Partition function of Kaur and Rana” studies p(λ)p(\lambda)9, where the largest part IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)00 appears exactly once and the remaining parts form a partition of IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)01, giving

IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)02

The paper then introduces IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)03, IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)04, IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)05, IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)06, IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)07, IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)08, IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)09, IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)10, IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)11, and IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)12 by imposing IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)13-regularity, overpartition, parity, color, cubic, or order restrictions on the partition of IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)14 (Johnson et al., 11 Oct 2025).

Continued-fraction methods provide a further representational extension. “Continued Fractions for partition generating functions” derives Euler-type continued fractions for products IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)15 and IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)16 and applies them to unrestricted, distinct, odd, binary, and IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)17-ary partition generating functions; it also uses Ramanujan’s techniques, including the Rogers–Ramanujan continued fraction and its two-parameter generalization, to connect continued fractions with product expansions and restricted partition families (Campbell, 2023).

5. Statistical-mechanical, graphical, and algebraic extensions

The paper “The Partition Function of Multicomponent Log-Gases” extends the one-dimensional log-gas from a single charge species to IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)18 species with integer charges IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)19 and fixed total charge

IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)20

The grand-canonical partition function is

IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)21

and, for IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)22 with IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)23 even and at most one odd IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)24, it admits the Berezin-integral representation

IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)25

where IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)26 is a non-homogeneous alternating tensor built from Wronskians of a complete polynomial family. This extends the de Bruijn IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)27 and IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)28 identities from single-species ensembles to multicomponent ensembles (Sinclair, 2011).

In graph theory, “Expansions of the Potts model partition function along deletions and contractions” extends the multivariate Tutte/Potts partition function

IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)29

from graphs to matroids by introducing the normalized form

IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)30

The paper proves deletion and contraction expansions,

IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)31

IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)32

and shows that they are dual under matroid duality, thereby extending chromatic-flow duality from planar graphs to the matroidal setting (Takahashi, 2024).

For the Ising model, “Graph theory and Pfaffian representations of Ising partition function” extends the usual planar Pfaffian picture to arbitrary embeddings by expressing the closed-curve partition function as the real part of the Pfaffian of a single skew-symmetric matrix with coefficients in a multicomplex algebra IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)33, where IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)34 is the non-orientable genus. The resulting formula is

IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)35

Known sums of IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)36 complex Pfaffians or IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)37 real Pfaffians are recovered by algebraic reduction. In this setting, the “extension” is topological and algebraic: the coefficients of the Pfaffian matrix are lifted from IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)38 to a multicomplex algebra that records the crosscap structure of the embedding surface (Gobron, 2013).

6. Computational extensions for temperature-dependent effective potentials

A computationally distinct meaning of the term appears in “Probing the partition function for temperature-dependent potentials with nested sampling.” Standard nested sampling yields the density of states in a single run when the potential IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)39 is temperature independent, but this property is lost when the effective potential depends explicitly on temperature,

IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)40

The paper restores the single-run strategy by introducing an extended partition function in which temperature is treated as an additional sampled variable. Operationally, the method samples joint points IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)41 with nested-sampling weights IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)42, then reconstructs thermodynamic quantities at a target IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)43 by a kernel IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)44. The internal-energy estimator is

IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)45

and the heat-capacity estimator contains the corresponding quadratic term in IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)46 together with the derivative correction involving IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)47. The paper applies this scheme to harmonic systems and to quantum Lennard-Jones clusters in the path-integral representation, where the ring-polymer effective potential is explicitly IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)48 dependent. Empirically, it reports that a uniform prior on IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)49 yields more balanced sampling than a uniform prior on IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)50, that IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)51 is preferred for IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)52 while IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)53 is preferred for IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)54, that the extended method recovers the direct method on IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)55 for IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)56 with appropriate IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)57, and that low-temperature sampling remains delicate for larger IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)58 and for IEext=(Ac+Ae)/(4G)I_E^{\rm ext}=-(A_c+A_e)/(4G)59 (Maillard et al., 2 Sep 2025).

Taken together, these usages indicate that an extended partition function is best understood as a family of extensions rather than a single theory. The extension may be ensemble-theoretic, geometric, algebraic, topological, combinatorial, or computational; what unifies the term is the systematic enlargement of a baseline partition function so that new observables, new constraints, or new structures become accessible.

Topic to Video (Beta)

No one has generated a video about this topic yet.

Whiteboard

No one has generated a whiteboard explanation for this topic yet.

Follow Topic

Get notified by email when new papers are published related to Extended Partition Function.