Marchesini–Onofri Equation Overview

• Marchesini–Onofri Equation is a singular integral eigenvalue equation characterizing adjoint excitations in large-N matrix quantum mechanics and connecting to variational and geometric frameworks.
• It is analyzed through numerical discretization, semiclassical Bohr–Sommerfeld quantization, and eigenfunction reformulations to verify universal Regge scaling and spectral transitions.
• The equation extends to diverse settings, including sharp exponential-integrability inequalities in geometric analysis and links to QCD dipole evolution in perturbative studies.

Searching arXiv for recent and foundational papers on the Marchesini–Onofri equation and closely related Onofri/Moser–Trudinger formulations. The Marchesini–Onofri equation denotes a cluster of closely related objects rather than a single universally normalized formula. In large-$N$ matrix quantum mechanics, it is a singular integral eigenvalue equation governing adjoint-sector excitations above the singlet Fermi sea. In the Onofri or Moser–Trudinger–Onofri variational framework, it is the Euler–Lagrange equation associated with sharp exponential-integrability inequalities, and in a recent upper-half-space formulation it appears as a quasi-linear Liouville equation with Neumann boundary data. A separate but structurally related QCD lineage connects Banfi–Marchesini–Smye evolution to the broader Marchesini–Onofri / Marchesini–Webber dipole-cascade picture, so the name also indexes a common dipole-evolution architecture across matrix models, geometric analysis, and perturbative gauge theory (Klebanov et al., 4 Mar 2026, Dolbeault et al., 2014, Dou et al., 21 Sep 2025, Hatta et al., 2013).

1. Adjoint-sector integral equation in matrix quantum mechanics

In the matrix-model usage, the setting is large-$N$ SU$(N)$-symmetric quantum mechanics of a Hermitian matrix $X$ with potential $V(X)$. The singlet sector is exactly solvable, with eigenvalues behaving as noninteracting fermions in $V$, and the ground state is obtained by filling the lowest $N$ levels up to a Fermi energy $\mu_F$. The adjoint sector is the first nontrivial SU$(N)$ representation, and its spectrum is organized as

$$E_{(s,n)}^{\rm adjoint}(g)=E_s^{\rm singlet}(g)+\Delta_n(g),\qquad n=1,2,\dots,$$

where the gaps $N$0 are universal in the sense that they do not depend on the singlet level $N$1 from which one starts (Klebanov et al., 4 Mar 2026).

The Marchesini–Onofri equation in this setting is the singular integral equation

$N$2

Here $N$3 are the classical turning points of the filled Fermi sea, $N$4 is the singlet eigenvalue density, $N$5 denotes principal value, and $N$6 is the adjoint excitation energy. The eigenfunctions are orthonormal with respect to the measure $N$7,

$N$8

and the adjoint constraint removes the constant mode,

$N$9

A useful reformulation introduces

$(N)$0

which converts the singular integral equation into a Schrödinger-like eigenvalue problem

$(N)$1

with

$(N)$2

This Hamiltonian form is the basis for both numerical and semiclassical analysis. A recurrent source of confusion is to identify this integral equation with the geometric Onofri Euler–Lagrange equation; the cited works instead use the same name for distinct but conceptually related structures.

2. Critical limit, time-of-flight variables, and spectral regimes

A key simplification is the time-of-flight coordinate

$(N)$3

which exposes the spectral structure near criticality. For the quartic potential

$(N)$4

the critical coupling sends the $(N)$5-domain to

$(N)$6

while for the cubic model

$(N)$7

the range is $(N)$8. The double-well model

$(N)$9

admits the same Marchesini–Onofri framework and displays analogous short-string and long-string regimes (Klebanov et al., 4 Mar 2026).

Near criticality, the low-lying spectrum obeys a Regge-like law

$X$0

For the quartic model, the paper gives

$X$1

with corresponding formulas for cubic and double-well potentials. The semiclassical interpretation treats the fold tip as a massless particle in a linear potential,

$X$2

so Bohr–Sommerfeld quantization,

$X$3

yields the Regge scaling. The same work states that the exact solution of the simplified Hamiltonian can be written in terms of Fresnel integrals, and that the large-$X$4 asymptotics correspond to Bohr–Sommerfeld with Maslov index $X$5 (Klebanov et al., 4 Mar 2026).

At sufficiently high excitation number, the spectrum crosses over to a linear WKB regime,

$X$6

This regime probes the full allowed $X$7-interval and is interpreted as a long-string sector. The low-lying Regge tower corresponds instead to short folded open strings. Slightly away from criticality, highly excited states become long strings extending far into the Liouville direction. The paper emphasizes that the Regge behavior is essentially universal, controlled by the critical near-top behavior and the string scale $X$8, while subleading corrections depend on the UV wall and on the specific potential.

3. Numerical treatment and universality claims

The adjoint Marchesini–Onofri equation is solved numerically by discretization on a grid. In the quartic case, after rescaling $X$9 and introducing a parameter $V(X)$0, the discretization uses

$V(X)$1

which yields a matrix eigenvalue problem

$V(X)$2

For numerical stability, the kernel is conjugated to a Hermitian matrix,

$V(X)$3

and the lowest eigenvalues are extracted with the Lanczos algorithm. A large-$V(X)$4 extrapolation is then performed,

$V(X)$5

The reported numerical results show that the first few quartic and cubic eigenvalues at $V(X)$6 agree extremely well with the analytic Regge formulas, including for $V(X)$7. The tabulated relative errors are stated to lie between the $V(X)$8 and $V(X)$9 level depending on $V$0, improving rapidly with excitation number. The same analysis also notes that resolving the Regge tower requires very small $V$1 and large matrix sizes $V$2, because the eigenfunctions become highly oscillatory at large $V$3 (Klebanov et al., 4 Mar 2026).

One specific conclusion is that the smallest gaps do not grow logarithmically with $V$4. The work highlights

$V$5

for the quartic theory near criticality and presents this as evidence against earlier conjectures of logarithmic growth. In the string-theoretic reading, this supports the statement that the singlet sector dominates the low-energy dynamics near criticality.

4. Euler–Lagrange equations in the Onofri and Moser–Trudinger framework

In geometric analysis, the Marchesini–Onofri equation is the stationarity equation for the Moser–Trudinger–Onofri functional. On a smooth, compact, connected Riemannian manifold $V$6 without boundary and normalized by $V$7, the inequality is written as

$V$8

In dimension $V$9, the relevant Euler–Lagrange equation is

$N$0

The paper treats this as a Marchesini–Onofri equation and proves a rigidity theorem: if $N$1 and $N$2, then every smooth solution is constant whenever $N$3, where

$N$4

The same work introduces the nonlinear flow

$N$5

under which the functional

$N$6

is monotone decreasing and satisfies an integral remainder estimate (Dolbeault et al., 2014).

The same variational scheme extends to weighted Euclidean space. For the stereographic weight

$N$7

the Euclidean Onofri inequality becomes

$N$8

with associated Euler–Lagrange equation

$N$9

A weighted rigidity threshold is defined by

$\mu_F$0

and for the stereographic weight one has $\mu_F$1, hence $\mu_F$2. The paper also notes the relevance of these weighted inequalities to the Keller–Segel chemotaxis model (Dolbeault et al., 2014).

A singular extension is developed for compact surfaces with conical data. The singular Liouville equation

$\mu_F$3

is transformed by

$\mu_F$4

into a weighted mean-field equation with

$\mu_F$5

The corresponding functional is

$\mu_F$6

In this singular Onofri-type setting, the sharp coefficient is altered by the strongest singularity through $\mu_F$7, and the sphere admits explicit sharp inequalities for one singularity and for two antipodal singularities, including an equality case when the antipodal singularities have equal negative strength (Mancini, 2014).

5. Upper-half-space trace analogue and boundary Liouville formulation

A recent higher-dimensional extension formulates an upper-half-space Marchesini–Onofri-type equation through a sharp Onofri trace inequality on

$\mu_F$8

The natural functional setting is a weighted Sobolev space $\mu_F$9 built from the conformal weight $(N)$0. The sharp trace inequality is

$(N)$1

where

$(N)$2

is sharp. Equality holds exactly for a translated and dilated bubble family, described in the paper as the trace or half-space analogues of standard Möbius bubbles (Dou et al., 21 Sep 2025).

The Euler–Lagrange equation is

$(N)$3

with

$(N)$4

After the transformation

$(N)$5

this becomes the quasi-linear Liouville-type Neumann problem

$(N)$6

The paper explicitly identifies this as the half-space Marchesini–Onofri-type equation.

The classification theorem states that every weak solution satisfying the finite-mass condition

$(N)$7

must be a translated and dilated bubble,

$(N)$8

for some $(N)$9 and $$E_{(s,n)}^{\rm adjoint}(g)=E_s^{\rm singlet}(g)+\Delta_n(g),\qquad n=1,2,\dots,$$0. Substituting back $$E_{(s,n)}^{\rm adjoint}(g)=E_s^{\rm singlet}(g)+\Delta_n(g),\qquad n=1,2,\dots,$$1 recovers the extremals of the trace inequality. The proof relies on a quasilinear Serrin–Zou-type identity, a Pohozaev-type identity, regularity estimates yielding

$$E_{(s,n)}^{\rm adjoint}(g)=E_s^{\rm singlet}(g)+\Delta_n(g),\qquad n=1,2,\dots,$$2

and asymptotics forcing the coefficient in

$$E_{(s,n)}^{\rm adjoint}(g)=E_s^{\rm singlet}(g)+\Delta_n(g),\qquad n=1,2,\dots,$$3

to satisfy $$E_{(s,n)}^{\rm adjoint}(g)=E_s^{\rm singlet}(g)+\Delta_n(g),\qquad n=1,2,\dots,$$4 (Dou et al., 21 Sep 2025).

6. Dipole evolution, non-global logarithms, and the QCD connection

The Marchesini–Onofri name also appears indirectly in perturbative QCD through its relation to color-dipole evolution. In the ATLAS jet-veto study, the central resummation tool is the Banfi–Marchesini–Smye equation, used to constrain inter-jet radiation for the veto fraction

$$E_{(s,n)}^{\rm adjoint}(g)=E_s^{\rm singlet}(g)+\Delta_n(g),\qquad n=1,2,\dots,$$5

The observable concerns dijet events at $$E_{(s,n)}^{\rm adjoint}(g)=E_s^{\rm singlet}(g)+\Delta_n(g),\qquad n=1,2,\dots,$$6 TeV with no additional jet above $$E_{(s,n)}^{\rm adjoint}(g)=E_s^{\rm singlet}(g)+\Delta_n(g),\qquad n=1,2,\dots,$$7 GeV in the rapidity interval between the primary jets, using anti-$$E_{(s,n)}^{\rm adjoint}(g)=E_s^{\rm singlet}(g)+\Delta_n(g),\qquad n=1,2,\dots,$$8 jets with $$E_{(s,n)}^{\rm adjoint}(g)=E_s^{\rm singlet}(g)+\Delta_n(g),\qquad n=1,2,\dots,$$9. In the regime $N$00, large logarithms of the form

$N$01

appear, and for large $N$02 one also encounters

$N$03

Fixed-order perturbation theory through $N$04 becomes unstable already for $N$05, motivating resummation (Hatta et al., 2013).

The structural link to the Marchesini–Onofri lineage is explicit. The paper states that

$N$06

and describes this as mathematically identical to the gluon-splitting structure in BFKL-like evolution. In that sense, the BMS equation belongs to the same broad dipole-cascade framework as Marchesini–Onofri and Marchesini–Webber evolution. It resums all leading single logarithms in the large-$N$07 approximation, including both Sudakov and non-global contributions. For the ATLAS observable, the resummed prediction is reported to give a good description of the data without fitted parameters, with non-global logarithms numerically less important than Sudakov suppression and reducing the gap fraction by about $N$08 at $N$09 GeV. The agreement is stated to be better for the most forward/backward jet-pair definition than for the two-leading-jet definition (Hatta et al., 2013).

This QCD application clarifies a useful distinction. The BMS equation is not the Marchesini–Onofri equation in the matrix-model or Onofri-functional sense. The relation is instead one of common dipole-evolution structure: nonlinear evolution of no-emission probabilities or adjoint excitations, real-virtual interplay, and a natural emergence of nonlocal kernels. That shared structure explains why the Marchesini–Onofri name persists across otherwise disparate domains.