Nehari–Pohozaev Manifold in Nonlinear PDEs

• The Nehari–Pohozaev manifold is a variational constraint that merges the Nehari condition with the Pohozaev identity to identify nontrivial critical points in complex PDEs.
• It utilizes specific fiber-map and scaling analyses to ensure unique projection and mountain-pass geometry, enhancing solution compactness and coercivity.
• This framework is instrumental in treating nonlocal and fractional problems, guaranteeing solution regularity, positivity, and precise bifurcation behavior.

Searching arXiv for recent and foundational papers on the Nehari–Pohozaev manifold. The Nehari–Pohozaev manifold is a variational constraint set defined by combining the Nehari condition, which encodes homogeneity along amplitude directions, with a Pohozaev identity, which encodes scaling balance for solutions of a given PDE. In contemporary nonlinear analysis it is used to isolate nontrivial critical points of energy functionals, especially for nonlocal, fractional, or zero-mass problems where the standard Nehari manifold alone may be insufficient. In the two sources considered here, it appears in two distinct but structurally related forms: for a two-component Schrödinger–Hartree–Fock system in $\mathbb{R}^2$ with logarithmic interactions (Albuquerque et al., 14 Nov 2025), and for the pseudo-relativistic equation $(-\Delta+m^2)^s u=f(u)$ in $\mathbb{R}^N$ with $0Bueno et al., 2018).

1. Variational origin and purpose

The central role of the Nehari–Pohozaev manifold is to encode, within a constrained variational framework, the identities that genuine solutions must satisfy. In the Hartree–Fock setting, the energy functional is defined on

$E:=W^\lambda\times W^\lambda,$

where

$$W^{\lambda} := \bigl\{\,u\in H^1(\mathbb{R}^2)\;:\;\|u\|_{ \lambda,*}^2:=\int_{\mathbb{R}^2}\log(\lambda+|x|)\,u^2(x)\,dx<\infty\bigr\},$$

with $\lambda>e^{1/4}$ and norm

$$\|u\|_{W^\lambda}^2=\|\nabla u\|_2^2+\|u\|_2^2+\|u\|_{\lambda,*}^2.$$

The functional is

$$I_\beta(u,v) = \frac12\Bigl(\|\nabla u\|_2^2+\|\nabla v\|_2^2\Bigr) +\frac14\,\mathcal V(u,v) -\frac1{2p}\,\psi_\beta(u,v),$$

with logarithmic interaction term $\mathcal V(u,v)$ and nonlinear term $(-\Delta+m^2)^s u=f(u)$0 as specified in the source (Albuquerque et al., 14 Nov 2025).

In the pseudo-relativistic setting, the variational structure is built on $(-\Delta+m^2)^s u=f(u)$1 with Fourier norm

$(-\Delta+m^2)^s u=f(u)$2

and energy

$(-\Delta+m^2)^s u=f(u)$3

where $(-\Delta+m^2)^s u=f(u)$4 (Bueno et al., 2018).

In both cases, the manifold serves as a “natural constraint” in the standard variational sense: critical points of the functional restricted to the manifold are genuine critical points in the ambient space. This is not merely a technical reformulation. It is the mechanism by which constrained minimization becomes compatible with the Euler–Lagrange equation and with the PDE’s scaling structure.

2. Nehari and Pohozaev constraints

The Nehari condition arises by testing the first variation against the unknown itself. For the Hartree–Fock system, this gives

$(-\Delta+m^2)^s u=f(u)$5

The same source states that any $(-\Delta+m^2)^s u=f(u)$6 solution satisfies the Pohozaev identity

$(-\Delta+m^2)^s u=f(u)$7

These are then combined into the functional

$(-\Delta+m^2)^s u=f(u)$8

with algebraic relation

$(-\Delta+m^2)^s u=f(u)$9

The resulting manifold is

$\mathbb{R}^N$0

Thus, in this formulation, the Nehari and Pohozaev data are merged into a single codimension-one constraint (Albuquerque et al., 14 Nov 2025).

By contrast, in the pseudo-relativistic problem the two constraints are imposed explicitly and simultaneously. The Nehari constraint is

$\mathbb{R}^N$1

that is,

$\mathbb{R}^N$2

while the Pohozaev functional is defined by

$\mathbb{R}^N$3

The manifold is then

$\mathbb{R}^N$4

This difference in presentation is significant. The Hartree–Fock paper packages the two balances into one scalar equation; the pseudo-relativistic paper keeps them separate. A plausible implication is that the optimal formulation depends on the scaling algebra of the model and on how many independent balance laws are needed to control the variational geometry.

3. Manifold structure and natural-constraint property

For the Hartree–Fock system, the set

$\mathbb{R}^N$5

is proved to be a $\mathbb{R}^N$6 submanifold of codimension $\mathbb{R}^N$7 in $\mathbb{R}^N$8 (Albuquerque et al., 14 Nov 2025). The key transversality condition is

$\mathbb{R}^N$9

and the source states that a direct calculation, using $0

$0

The implicit-function theorem then applies.

The same strict transversality enters the natural-constraint lemma: if $0minimizer together with the inequality $0Albuquerque et al., 14 Nov 2025).

In the pseudo-relativistic problem, the source describes the manifold as a natural closed manifold on which $0Bueno et al., 2018). It likewise asserts that continuity of the projection map supplied by fiber analysis yields a Lagrange-multiplier characterization of critical points of $0$E:=W^\lambda\times W^\lambda,$0 are true solutions.

A recurring misconception is that the Nehari–Pohozaev manifold is simply a decorative strengthening of the Nehari manifold. The two sources indicate otherwise. In both settings, the Pohozaev information is structurally active: it helps recover coercivity, excludes incorrect scalings, and aligns the constrained variational problem with the actual solution class.

4. Fibering maps and scaling geometry

The geometry of the manifold is governed by a one-parameter scaling analysis. In the Hartree–Fock problem, for each nonzero $E:=W^\lambda\times W^\lambda,$1 one defines

$E:=W^\lambda\times W^\lambda,$2

and studies

$E:=W^\lambda\times W^\lambda,$3

The source states that

$E:=W^\lambda\times W^\lambda,$4

hence

$E:=W^\lambda\times W^\lambda,$5

Therefore the critical points of the scalar fiber map correspond exactly to points on the Nehari–Pohozaev manifold. Moreover,

$E:=W^\lambda\times W^\lambda,$6

It follows that $E:=W^\lambda\times W^\lambda,$7 has a unique maximum $E:=W^\lambda\times W^\lambda,$8 and $E:=W^\lambda\times W^\lambda,$9. Each nonzero pair is therefore projected exactly once onto $$W^{\lambda} := \bigl\{\,u\in H^1(\mathbb{R}^2)\;:\;\|u\|_{ \lambda,*}^2:=\int_{\mathbb{R}^2}\log(\lambda+|x|)\,u^2(x)\,dx<\infty\bigr\},$$0, and the restriction of the energy inherits a mountain-pass-type geometry along each such ray (Albuquerque et al., 14 Nov 2025).

In the pseudo-relativistic problem, the fiber map is the simpler amplitude scaling

$$W^{\lambda} := \bigl\{\,u\in H^1(\mathbb{R}^2)\;:\;\|u\|_{ \lambda,*}^2:=\int_{\mathbb{R}^2}\log(\lambda+|x|)\,u^2(x)\,dx<\infty\bigr\},$$1

The derivative satisfies

$$W^{\lambda} := \bigl\{\,u\in H^1(\mathbb{R}^2)\;:\;\|u\|_{ \lambda,*}^2:=\int_{\mathbb{R}^2}\log(\lambda+|x|)\,u^2(x)\,dx<\infty\bigr\},$$2

Under the hypotheses $$W^{\lambda} := \bigl\{\,u\in H^1(\mathbb{R}^2)\;:\;\|u\|_{ \lambda,*}^2:=\int_{\mathbb{R}^2}\log(\lambda+|x|)\,u^2(x)\,dx<\infty\bigr\},$$3 and $$W^{\lambda} := \bigl\{\,u\in H^1(\mathbb{R}^2)\;:\;\|u\|_{ \lambda,*}^2:=\int_{\mathbb{R}^2}\log(\lambda+|x|)\,u^2(x)\,dx<\infty\bigr\},$$4 quoted in the source, one has

$$W^{\lambda} := \bigl\{\,u\in H^1(\mathbb{R}^2)\;:\;\|u\|_{ \lambda,*}^2:=\int_{\mathbb{R}^2}\log(\lambda+|x|)\,u^2(x)\,dx<\infty\bigr\},$$5

and there exists a unique $$W^{\lambda} := \bigl\{\,u\in H^1(\mathbb{R}^2)\;:\;\|u\|_{ \lambda,*}^2:=\int_{\mathbb{R}^2}\log(\lambda+|x|)\,u^2(x)\,dx<\infty\bigr\},$$6 at which $$W^{\lambda} := \bigl\{\,u\in H^1(\mathbb{R}^2)\;:\;\|u\|_{ \lambda,*}^2:=\int_{\mathbb{R}^2}\log(\lambda+|x|)\,u^2(x)\,dx<\infty\bigr\},$$7 attains its global maximum; $$W^{\lambda} := \bigl\{\,u\in H^1(\mathbb{R}^2)\;:\;\|u\|_{ \lambda,*}^2:=\int_{\mathbb{R}^2}\log(\lambda+|x|)\,u^2(x)\,dx<\infty\bigr\},$$8 is strictly increasing before $$W^{\lambda} := \bigl\{\,u\in H^1(\mathbb{R}^2)\;:\;\|u\|_{ \lambda,*}^2:=\int_{\mathbb{R}^2}\log(\lambda+|x|)\,u^2(x)\,dx<\infty\bigr\},$$9 and strictly decreasing afterward (Bueno et al., 2018).

These fiber constructions are not identical. The Hartree–Fock model uses a Pohozaev scaling adapted to the logarithmic interaction in $\lambda>e^{1/4}$0, whereas the pseudo-relativistic model uses ordinary amplitude scaling. This suggests that the term “Nehari–Pohozaev manifold” designates a method rather than a single canonical formula: the constraint is built to track the scaling law relevant to the PDE.

5. Ground states and compactness mechanisms

The Hartree–Fock paper defines the variational level

$\lambda>e^{1/4}$1

and proves that if $\lambda>e^{1/4}$2, $\lambda>e^{1/4}$3, and $\lambda>e^{1/4}$4, then there exists a nonnegative ground state $\lambda>e^{1/4}$5 such that

$\lambda>e^{1/4}$6

The proof uses boundedness of minimizing sequences, avoidance of vanishing via Lions’ lemma plus translations, and Palais–Smale compactness on $\lambda>e^{1/4}$7 (Albuquerque et al., 14 Nov 2025).

The same theorem records asymptotic behavior in the interaction parameter: as $\lambda>e^{1/4}$8,

$\lambda>e^{1/4}$9

up to subsequence; and as $$\|u\|_{W^\lambda}^2=\|\nabla u\|_2^2+\|u\|_2^2+\|u\|_{\lambda,*}^2.$$0,

$$\|u\|_{W^\lambda}^2=\|\nabla u\|_2^2+\|u\|_2^2+\|u\|_{\lambda,*}^2.$$1

These statements show that the constrained variational framework is stable enough to capture limiting regimes of the coupling parameter (Albuquerque et al., 14 Nov 2025).

For the pseudo-relativistic equation, under assumptions $$\|u\|_{W^\lambda}^2=\|\nabla u\|_2^2+\|u\|_2^2+\|u\|_{\lambda,*}^2.$$2–$$\|u\|_{W^\lambda}^2=\|\nabla u\|_2^2+\|u\|_2^2+\|u\|_{\lambda,*}^2.$$3, and in particular for

$$\|u\|_{W^\lambda}^2=\|\nabla u\|_2^2+\|u\|_2^2+\|u\|_{\lambda,*}^2.$$4

the source states that there exists a nontrivial ground-state solution $$\|u\|_{W^\lambda}^2=\|\nabla u\|_2^2+\|u\|_2^2+\|u\|_{\lambda,*}^2.$$5 minimizing $$\|u\|_{W^\lambda}^2=\|\nabla u\|_2^2+\|u\|_2^2+\|u\|_{\lambda,*}^2.$$6 on $$\|u\|_{W^\lambda}^2=\|\nabla u\|_2^2+\|u\|_2^2+\|u\|_{\lambda,*}^2.$$7:

$$\|u\|_{W^\lambda}^2=\|\nabla u\|_2^2+\|u\|_2^2+\|u\|_{\lambda,*}^2.$$8

The proof uses a variant of the Ghoussoub–Preiss theorem to produce a Cerami sequence,

$$\|u\|_{W^\lambda}^2=\|\nabla u\|_2^2+\|u\|_2^2+\|u\|_{\lambda,*}^2.$$9

followed by concentration–compactness à la Lions and a translation-invariance argument to preclude vanishing and dichotomy (Bueno et al., 2018).

The two papers illustrate a standard pattern in the subject: once the manifold is chosen so that every nonzero configuration admits a unique projection onto it, compactness can be studied at the projected level, where spurious scaling freedom has been removed.

6. Regularity, positivity, and qualitative bifurcation

Beyond existence, the Hartree–Fock analysis establishes qualitative properties of solutions. Every weak solution in $$I_\beta(u,v) = \frac12\Bigl(\|\nabla u\|_2^2+\|\nabla v\|_2^2\Bigr) +\frac14\,\mathcal V(u,v) -\frac1{2p}\,\psi_\beta(u,v),$$0 is in fact in $$I_\beta(u,v) = \frac12\Bigl(\|\nabla u\|_2^2+\|\nabla v\|_2^2\Bigr) +\frac14\,\mathcal V(u,v) -\frac1{2p}\,\psi_\beta(u,v),$$1 and is strictly positive in each nontrivial component. The Pohozaev identity

$$I_\beta(u,v) = \frac12\Bigl(\|\nabla u\|_2^2+\|\nabla v\|_2^2\Bigr) +\frac14\,\mathcal V(u,v) -\frac1{2p}\,\psi_\beta(u,v),$$2

is also rigorously derived for weak solutions, confirming that the identity used to define the constraint is not merely formal (Albuquerque et al., 14 Nov 2025).

The ground-state structure undergoes a coupling-dependent transition. Writing the ground state as $$I_\beta(u,v) = \frac12\Bigl(\|\nabla u\|_2^2+\|\nabla v\|_2^2\Bigr) +\frac14\,\mathcal V(u,v) -\frac1{2p}\,\psi_\beta(u,v),$$3, the source states, via a suitable polarization or “mixing” argument, that:

Regime of $$I_\beta(u,v) = \frac12\Bigl(\|\nabla u\|_2^2+\|\nabla v\|_2^2\Bigr) +\frac14\,\mathcal V(u,v) -\frac1{2p}\,\psi_\beta(u,v),$$4	Ground-state type
$$I_\beta(u,v) = \frac12\Bigl(\|\nabla u\|_2^2+\|\nabla v\|_2^2\Bigr) +\frac14\,\mathcal V(u,v) -\frac1{2p}\,\psi_\beta(u,v),$$5	both components are nonzero (vector solution)
$$I_\beta(u,v) = \frac12\Bigl(\|\nabla u\|_2^2+\|\nabla v\|_2^2\Bigr) +\frac14\,\mathcal V(u,v) -\frac1{2p}\,\psi_\beta(u,v),$$6	one component vanishes (semi-trivial)
$$I_\beta(u,v) = \frac12\Bigl(\|\nabla u\|_2^2+\|\nabla v\|_2^2\Bigr) +\frac14\,\mathcal V(u,v) -\frac1{2p}\,\psi_\beta(u,v),$$7	$$I_\beta(u,v) = \frac12\Bigl(\|\nabla u\|_2^2+\|\nabla v\|_2^2\Bigr) +\frac14\,\mathcal V(u,v) -\frac1{2p}\,\psi_\beta(u,v),$$8 solves the symmetric decoupled problem

This bifurcation picture is one of the sharper uses of the Nehari–Pohozaev framework in the source: the manifold is not only a vehicle for existence but also a setting in which the competition between vector and semi-trivial states can be characterized precisely (Albuquerque et al., 14 Nov 2025).

In the pseudo-relativistic paper, the abstract states that positive solutions are radially symmetric and decreasing with respect to the origin for classes of nonlinearities modeled by functions like $$I_\beta(u,v) = \frac12\Bigl(\|\nabla u\|_2^2+\|\nabla v\|_2^2\Bigr) +\frac14\,\mathcal V(u,v) -\frac1{2p}\,\psi_\beta(u,v),$$9, $\mathcal V(u,v)$0, or $\mathcal V(u,v)$1 (Bueno et al., 2018). Although that symmetry result is not presented in the detailed summary as a direct consequence of the Nehari–Pohozaev manifold, it situates the manifold method within a broader qualitative theory where variational existence is complemented by symmetry and monotonicity.

7. Relation to Pohozaev identities and broader methodological significance

The manifold is inseparable from the Pohozaev identity that motivates it. In the pseudo-relativistic problem, the source first proves a Pohozaev-type identity for the extension problem in $\mathcal V(u,v)$2 and then translates it to the original equation, obtaining

$\mathcal V(u,v)$3

That identity is then built directly into the constraint set (Bueno et al., 2018).

In the Hartree–Fock system, the analogous Pohozaev identity reflects the balance between kinetic energy, the logarithmic Hartree interaction, the $\mathcal V(u,v)$4 mass term, and the coupled power nonlinearity. The combined constraint

$\mathcal V(u,v)$5

shows explicitly how Nehari and Pohozaev balances are fused into a single scalar equation (Albuquerque et al., 14 Nov 2025).

Taken together, these examples show that the Nehari–Pohozaev manifold is especially useful in three situations documented by the sources. First, it is effective when the PDE has a nontrivial scaling law that is not adequately captured by the standard Nehari condition alone. Second, it is useful in nonlocal settings, including logarithmic Hartree interactions and fractional pseudo-relativistic operators. Third, it provides a framework in which constrained minimization, fiber-map uniqueness, and concentration–compactness can be combined without losing contact with the exact identities satisfied by solutions.

A plausible implication is that the method’s durability comes from its adaptability: the manifold is not defined by one universal formula, but by the pairing of a variational energy with the specific Pohozaev balance of the underlying equation. In that sense, the Nehari–Pohozaev manifold is best understood as a structural principle in nonlinear PDE rather than a single fixed object.