Elliptic Ruijsenaars–Schneider Model

• The elliptic Ruijsenaars–Schneider model is an integrable many-body system that generalizes the Calogero–Moser model using elliptic functions and a relativistic deformation parameter.
• It features a Lax representation with a spectral curve that yields commuting Hamiltonians and connects with the Toda lattice hierarchy via pole dynamics.
• The quantum formulation employs elliptic difference operators and Q-operators, establishing duality structures and non-perturbative quantization links with gauge theories.

The elliptic Ruijsenaars–Schneider model is the elliptic, relativistic deformation of the Calogero–Moser many-body system and is commonly presented as an $N$-particle integrable system with canonical variables $(x_i,p_i)$, Poisson brackets $\{x_i,p_j\}=\delta_{ij}$, an elliptic interaction built from Weierstrass or theta functions, and a Lax representation with spectral parameter on an elliptic curve. In the formulation used in the hierarchy literature, it is the $N$-particle relativistic Calogero–Moser model governing the pole dynamics of elliptic solutions of the $2$D Toda lattice hierarchy; in the quantum setting it becomes an elliptic $q$-difference system of type $A$ with commuting Hamiltonians, exact quantization problems, Baxter-type $Q$-operators, and a nontrivial duality structure (Prokofev et al., 2021, Rains et al., 23 Mar 2025, Hatsuda et al., 2018, Mironov et al., 2021).

1. Canonical formulation and equations of motion

In the classical elliptic RS model used in the Toda correspondence, the commuting integrals are

$I_k=\sum_{\substack{I\subset\{1,\dots,N\}\ |I|=k} \exp\!\left(\sum_{i\in I}p_i\right) \prod_{\substack{i\in I\ j\notin I} } \frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad k=1,\dots,N,$

with

$$I_1=\sum_i e^{p_i}\prod_{j\neq i}\frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad I_N=e^{\sum_i p_i}.$$

The parameter $(x_i,p_i)$0 is the relativistic deformation parameter or lattice shift. The convention employed in this formulation differs from Ruijsenaars’ original one by a canonical transformation eliminating square roots (Prokofev et al., 2021).

The Hamiltonian $(x_i,p_i)$1 yields the velocity law

$(x_i,p_i)$2

and the second-order equations of motion

$(x_i,p_i)$3

Equivalently,

$(x_i,p_i)$4

The same second-order system is obtained from the $(x_i,p_i)$5- and $(x_i,p_i)$6-dynamics in the $(x_i,p_i)$7D Toda setting (Prokofev et al., 2021).

An equivalent elliptic presentation, used in the self-dual formulation, writes the equations as

$(x_i,p_i)$8

The functions here are defined through the odd theta-function $(x_i,p_i)$9, the Kronecker function

$\{x_i,p_j\}=\delta_{ij}$0

and the Eisenstein function

$\{x_i,p_j\}=\delta_{ij}$1

Their relation to the Weierstrass notation is

$\{x_i,p_j\}=\delta_{ij}$2

This makes explicit that the zeta-function and theta-function formulations are equivalent up to standard normalization choices (Zabrodin et al., 2017).

A recurrent source of confusion is that several normalizations coexist in the literature: sigma/zeta conventions emphasize the algebro-geometric and Toda-hierarchy structure, whereas theta/Kronecker conventions are adapted to self-duality, Bäcklund transformations, and quantum difference operators. The cited works use both, but they describe the same elliptic RS dynamics (Prokofev et al., 2021, Zabrodin et al., 2017).

2. Lax representation, spectral curve, and commuting Hamiltonians

The elliptic RS equations admit a Lax representation extracted from the Toda auxiliary linear problem. The basic Lax matrix is

$\{x_i,p_j\}=\delta_{ij}$3

or, in canonical variables,

$\{x_i,p_j\}=\delta_{ij}$4

where

$\{x_i,p_j\}=\delta_{ij}$5

The associated $\{x_i,p_j\}=\delta_{ij}$6-matrix satisfies the compatibility condition

$\{x_i,p_j\}=\delta_{ij}$7

and this compatibility is equivalent to the elliptic RS equations of motion (Prokofev et al., 2021).

The spectral curve is initially written as

$\{x_i,p_j\}=\delta_{ij}$8

Because $\{x_i,p_j\}=\delta_{ij}$9 has an essential singularity at $N$0, one introduces a gauge-transformed matrix

$N$1

with

$N$2

After setting $N$3, the spectral curve becomes

$N$4

Using the elliptic Cauchy identity, the determinant takes the explicit form

$N$5

so the coefficients of the spectral polynomial are precisely the commuting RS integrals $N$6 (Prokofev et al., 2021).

Two marked points play a distinguished role: $N$7 Near these points, the local expansions of $N$8 generate the higher Hamiltonians of the elliptic RS hierarchy: $N$9 and

$2$0

The corresponding flows are Hamiltonian: $2$1

$2$2

This identifies the full positive and negative Toda hierarchies with two commuting halves of the elliptic RS hierarchy generated by the same spectral curve (Prokofev et al., 2021).

In the rational and trigonometric limits, the same construction degenerates to the familiar hierarchy Hamiltonians expressed through $2$3 or $2$4, while in the $2$5 limit the positive Hamiltonians reduce to Calogero–Moser Hamiltonians and the negative ones disappear (Prokofev et al., 2021).

3. Elliptic Toda correspondence and pole dynamics

A central structural result is that the elliptic RS model arises as a pole reduction of the $2$6D Toda lattice hierarchy. The relevant class of tau-functions is

$2$7

with distinct zeros $2$8 in a fundamental domain. For such solutions, the Toda field

$2$9

takes the elliptic form

$q$0

The wave function and its adjoint are double-Bloch functions, and for elliptic solutions they admit the pole ansatz

$q$1

with spectral relation

$q$2

Substituting this ansatz into the Toda auxiliary linear problem yields the finite-dimensional linear system

$q$3

whose compatibility reproduces the elliptic RS dynamics (Prokofev et al., 2021).

The hierarchy-level extension is the main theorem of the pole-dynamics analysis. It does not stop at the classical statement that poles move according to the $q$4-flow of elliptic RS; rather, it proves that all Toda times $q$5 and $q$6 induce the higher commuting Hamiltonian flows of the elliptic RS hierarchy, with Hamiltonians read off from the two local expansions of the RS spectral curve at $q$7 and $q$8 (Prokofev et al., 2021).

This makes the elliptic RS model the finite-dimensional integrable system underlying elliptic $q$9-dependent solutions of the $A$0D Toda hierarchy. A plausible implication is that the spectral-curve viewpoint is more fundamental than the individual equations of motion, because it simultaneously organizes the Lax pair, commuting integrals, and the positive/negative Toda flows. That implication is consistent with the way the marked-point expansions generate the entire hierarchy (Prokofev et al., 2021).

The same pole-ansatz philosophy extends beyond the ordinary Toda hierarchy. For elliptic solutions of the Toda lattice with constraint of type $A$1, the pole dynamics is no longer the standard elliptic RS system but a deformed variant with an additional many-body force term. In that setting one has

$A$2

with

$A$3

At $A$4, this deformation reduces to the ordinary elliptic RS equations. The constrained-Toda origin also leads to a Manakov-triple representation and a spectral curve with involution $A$5 (Prokofev et al., 2023).

4. Self-duality, Bäcklund transformations, and field analogues

A different but closely related formulation is the self-dual or Bäcklund-type first-order system. In its theta-function form,

$A$6

$A$7

These equations imply the RS equations for the $A$8, while the $A$9 satisfy the $Q$0 RS equations. In the elliptic case one must have $Q$1, so the self-dual elliptic RS model involves two sets of $Q$2 complex variables, $Q$3 and $Q$4, obeying the same second-order dynamics and related by a first-order Bäcklund system (Zabrodin et al., 2017).

The elliptic case is special because the derivation uses double periodicity and elliptic identities that require equal numbers of $Q$5 and $Q$6 factors. An alternative derivation starts from the complexified ILW equation with discrete Laplacian,

$Q$7

with

$Q$8

and pole ansatz

$Q$9

Comparing residues yields the self-dual equations. This places elliptic RS in the same pole-dynamics lineage that relates Calogero–Moser systems to ILW and Benjamin–Ono equations (Zabrodin et al., 2017).

The field-theoretic extension replaces finitely many particle coordinates by fields $I_k=\sum_{\substack{I\subset\{1,\dots,N\}\ |I|=k} \exp\!\left(\sum_{i\in I}p_i\right) \prod_{\substack{i\in I\ j\notin I} } \frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad k=1,\dots,N,$0 depending on continuous time and continuous space, while retaining shifts by $I_k=\sum_{\substack{I\subset\{1,\dots,N\}\ |I|=k} \exp\!\left(\sum_{i\in I}p_i\right) \prod_{\substack{i\in I\ j\notin I} } \frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad k=1,\dots,N,$1 in the spatial variable. In one construction, a chain product of elliptic $I_k=\sum_{\substack{I\subset\{1,\dots,N\}\ |I|=k} \exp\!\left(\sum_{i\in I}p_i\right) \prod_{\substack{i\in I\ j\notin I} } \frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad k=1,\dots,N,$2-matrices yields a lattice RS chain whose Hamiltonian is

$I_k=\sum_{\substack{I\subset\{1,\dots,N\}\ |I|=k} \exp\!\left(\sum_{i\in I}p_i\right) \prod_{\substack{i\in I\ j\notin I} } \frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad k=1,\dots,N,$3

with normalized Lax matrix

$I_k=\sum_{\substack{I\subset\{1,\dots,N\}\ |I|=k} \exp\!\left(\sum_{i\in I}p_i\right) \prod_{\substack{i\in I\ j\notin I} } \frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad k=1,\dots,N,$4

In the second construction, elliptic families of $I_k=\sum_{\substack{I\subset\{1,\dots,N\}\ |I|=k} \exp\!\left(\sum_{i\in I}p_i\right) \prod_{\substack{i\in I\ j\notin I} } \frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad k=1,\dots,N,$5D Toda solutions produce a semi-discrete zero-curvature system

$I_k=\sum_{\substack{I\subset\{1,\dots,N\}\ |I|=k} \exp\!\left(\sum_{i\in I}p_i\right) \prod_{\substack{i\in I\ j\notin I} } \frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad k=1,\dots,N,$6

and field equations

$I_k=\sum_{\substack{I\subset\{1,\dots,N\}\ |I|=k} \exp\!\left(\sum_{i\in I}p_i\right) \prod_{\substack{i\in I\ j\notin I} } \frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad k=1,\dots,N,$7

$I_k=\sum_{\substack{I\subset\{1,\dots,N\}\ |I|=k} \exp\!\left(\sum_{i\in I}p_i\right) \prod_{\substack{i\in I\ j\notin I} } \frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad k=1,\dots,N,$8

Under the reduction $I_k=\sum_{\substack{I\subset\{1,\dots,N\}\ |I|=k} \exp\!\left(\sum_{i\in I}p_i\right) \prod_{\substack{i\in I\ j\notin I} } \frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad k=1,\dots,N,$9, this field model reproduces the ordinary elliptic RS system; in the limit $$I_1=\sum_i e^{p_i}\prod_{j\neq i}\frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad I_N=e^{\sum_i p_i}.$$0, it reduces to the field Calogero–Moser theory (Zabrodin et al., 2021).

These constructions suggest that the elliptic RS model is not merely a finite-dimensional many-body system but also a node in a broader network including Bäcklund systems, nonlocal PDEs, spin chains, and semi-discrete field theories. That suggestion is explicit in the cited formulations, though the precise categorical relation between all of these realizations is not developed uniformly across the papers (Zabrodin et al., 2017, Zabrodin et al., 2021).

5. Geometric realizations, reductions, and deformations

A different geometric direction concerns compact real forms. For type (i) couplings, the center-of-mass phase space of the trigonometric III$$I_1=\sum_i e^{p_i}\prod_{j\neq i}\frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad I_N=e^{\sum_i p_i}.$$1 and elliptic IV$$I_1=\sum_i e^{p_i}\prod_{j\neq i}\frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad I_N=e^{\sum_i p_i}.$$2 RS systems can be completed to $$I_1=\sum_i e^{p_i}\prod_{j\neq i}\frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad I_N=e^{\sum_i p_i}.$$3 with rescaled Fubini–Study form. The completion is achieved through an explicit symplectic embedding of the local phase space into projective space, and the local elliptic Lax matrix extends to a global smooth matrix

$$I_1=\sum_i e^{p_i}\prod_{j\neq i}\frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad I_N=e^{\sum_i p_i}.$$4

on $$I_1=\sum_i e^{p_i}\prod_{j\neq i}\frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad I_N=e^{\sum_i p_i}.$$5, satisfying

$$I_1=\sum_i e^{p_i}\prod_{j\neq i}\frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad I_N=e^{\sum_i p_i}.$$6

As a result, the coefficients of the characteristic polynomial of the global elliptic Lax matrix define globally smooth commuting Hamiltonians on $$I_1=\sum_i e^{p_i}\prod_{j\neq i}\frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad I_N=e^{\sum_i p_i}.$$7 (Feher et al., 2016).

Another direction is the type-$$I_1=\sum_i e^{p_i}\prod_{j\neq i}\frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad I_N=e^{\sum_i p_i}.$$8 deformation. The deformed elliptic RS system can be realized as a reduction of the ordinary elliptic RS model with an even number of particles to a subspace where particles form rigid pairs at distance $$I_1=\sum_i e^{p_i}\prod_{j\neq i}\frac{\sigma(x_i-x_j+\eta)}{\sigma(x_i-x_j)}, \qquad I_N=e^{\sum_i p_i}.$$9: $(x_i,p_i)$00 The relevant commuting family is the minus-half of the RS hierarchy,

$(x_i,p_i)$01

which preserves the pair subspace, whereas the plus-flows do not. Restricting the standard RS integrals to this invariant Lagrangian subspace yields explicit conserved quantities $(x_i,p_i)$02 for the deformed model and a generating determinant

$(x_i,p_i)$03

The resulting spectral curve has the involution

$(x_i,p_i)$04

reflecting the paired-particle origin of the reduction (Zabrodin, 2022).

The time-discretized version of this deformed system is obtained from Bäcklund transformations and from pole dynamics of elliptic solutions of the fully discrete BKP equation. The continuous-time deformed RS system differs from the ordinary elliptic RS equations by the additional term proportional to $(x_i,p_i)$05, while the discrete-time equations generalize the Nijhoff–Ragnisco–Kuznetsov map for standard RS. The paper presenting this discretization explicitly leaves open whether the discrete equations admit a commutation representation such as a discrete Lax, Manakov-triple, or Zakharov–Shabat form (Zabrodin, 2023).

A further generalization embeds the spin elliptic $(x_i,p_i)$06 RS system into a family of relativistic interacting elliptic $(x_i,p_i)$07 tops. In the $(x_i,p_i)$08 reduction, the block Lax matrix

$(x_i,p_i)$09

and companion $(x_i,p_i)$10-matrix recover the Krichever–Zabrodin spin elliptic RS model. The same paper states explicitly that the Poisson structure and classical $(x_i,p_i)$11-matrix structure for the spin elliptic RS model are unknown, while the trigonometric and rational cases are better understood (Zotov, 2019).

6. Quantum elliptic RS: difference operators, duality, $(x_i,p_i)$12-operators, and exact quantization

In the quantum theory, the elliptic RS model becomes a commuting family of difference operators. For the type-$(x_i,p_i)$13 quantum elliptic RS system, the higher Hamiltonians are

$(x_i,p_i)$14

with

$(x_i,p_i)$15

and Ruijsenaars’ theorem gives

$(x_i,p_i)$16

This is the genuinely elliptic $(x_i,p_i)$17 quantum RS system, depending on elliptic nomes $(x_i,p_i)$18 and coupling $(x_i,p_i)$19 (Rains et al., 23 Mar 2025).

A major recent advance is the construction of a one-parameter commuting family of $(x_i,p_i)$20-operators,

$(x_i,p_i)$21

which satisfy

$(x_i,p_i)$22

Their commutativity is equivalent to the symmetry of an elliptic hypergeometric kernel $(x_i,p_i)$23, and the same operators are simultaneously diagonal on the elliptic Macdonald basis of Langmann–Noumi–Shiraishi (Rains et al., 23 Mar 2025).

At special values of the spectral parameter, the $(x_i,p_i)$24-operators degenerate to Noumi–Sano operators. Precisely, $(x_i,p_i)$25 has at most a simple pole at $(x_i,p_i)$26, and its residue reproduces $(x_i,p_i)$27, the gauge-transformed Noumi–Sano difference operator. This gives a direct bridge between integral Baxter operators and the commuting elliptic difference Hamiltonians (Rains et al., 23 Mar 2025).

A complementary line of work concerns duality and eigenfunctions. The elliptic RS Hamiltonians in multiplicative variables $(x_i,p_i)$28 are

$(x_i,p_i)$29

and the work on elliptic symmetric functions identifies two distinct elliptic analogues of Macdonald polynomials. The functions $(x_i,p_i)$30 are eigenfunctions of the dual elliptic RS system, while $(x_i,p_i)$31 diagonalize an elliptic reduction of the Koroteev–Shakirov Hamiltonians. They are related by a Schur-orthogonal conjugation

$(x_i,p_i)$32

not by literal equality. In the trigonometric limit $(x_i,p_i)$33,

$(x_i,p_i)$34

so the usual Macdonald self-duality is recovered only after degeneration (Mironov et al., 2021).

Exact quantization of the quantum elliptic RS spectrum has been formulated through the correspondence with five-dimensional $(x_i,p_i)$35 $(x_i,p_i)$36 gauge theory in the Nekrasov–Shatashvili limit. For the $(x_i,p_i)$37-particle quantum eRS model on a torus with half-periods $(x_i,p_i)$38, the Hamiltonian is

$(x_i,p_i)$39

The paper distinguishes a B-model quantization, where the electric periods are quantized,

$(x_i,p_i)$40

and an A-model quantization, where one imposes exact dual-period conditions through the full twisted superpotential

$(x_i,p_i)$41

In the relativistic elliptic case, unlike the nonrelativistic elliptic Calogero–Moser case, the A-type quantization requires non-perturbative completion in $(x_i,p_i)$42, fixed by modular-duality symmetry $(x_i,p_i)$43 (Hatsuda et al., 2018).

Recent gauge-theoretic comparisons refine this picture. Superconformal indices with surface defects produce perturbative elliptic RS eigenfunctions, while ramified instanton partition functions in the NS limit produce formal RS eigenfunctions and Wilson-loop eigenvalues; after imposing the B-model exact quantization on Coulomb moduli, the two constructions agree for the low-lying $(x_i,p_i)$44 and $(x_i,p_i)$45 states that were checked explicitly (Kim et al., 2024). Physics-inspired large-compactification arguments have also yielded perturbative ground-state wavefunctions and eigenvalues for the elliptic $(x_i,p_i)$46 RS operator, though those derivations are presented as relying on physical input and perturbative checks rather than as closed-form spectral solutions (Nazzal et al., 2023).

One recent comparison with finite-volume sine-Gordon theory sharpens the role of the elliptic model as a periodic finite-volume analogue of hyperbolic RS. In the two-particle center-of-mass sector, the exact finite-volume spectrum of elliptic RS is not equal to that of finite-volume sine-Gordon. The same study finds that Bethe–Yang quantization based only on the infinite-volume scattering phase is exact in the rational and trigonometric cases, but acquires finite-size corrections in the hyperbolic and elliptic cases, both relativistically and in the non-relativistic Calogero–Moser limit (Bajnok et al., 20 Nov 2025).

Taken together, these results show that the elliptic Ruijsenaars–Schneider model is simultaneously a classical many-body system, a hierarchy reduction of Toda-type equations, a source of self-dual and field-theoretic extensions, and a quantum elliptic difference system with commuting Hamiltonians, spectral dualities, $(x_i,p_i)$47-operators, and non-perturbative exact quantization problems.