Keller–Osserman Transform Analysis

• The Keller–Osserman transform is a method that establishes existence or blow-up of solutions in nonlinear elliptic equations by linking the growth of reaction terms to an integral threshold.
• It is applied in diverse settings including entire-space problems, boundary blow-up, Hessian, and fractional cases, illustrating its versatility across different operators.
• Advanced formulations extend the transform to coupled systems and geometric contexts, providing scalar reductions that yield explicit asymptotic rates and uniqueness results.

Searching arXiv for recent and foundational papers on the Keller–Osserman transform and related Keller–Osserman conditions. The Keller–Osserman transform is an integral device in nonlinear elliptic theory that converts the growth of a reaction term into a scalar quantity governing existence, nonexistence, blow-up, and asymptotic behavior of solutions. In the classical semilinear equation $\Delta u=f(u)$, it is tied to the primitive $F(t)=\int_0^t f(s)\,ds$ and to the Keller–Osserman integral $\int_1^\infty \frac{dt}{\sqrt{F(t)}}$; in later developments, the same mechanism appears as an inverse boundary profile, as an operator-dependent transform involving $K^{-1}$, as a scalar reduction for coupled systems, and as an ODE for level-set measures when explicit radial barriers are unavailable (Sirakov et al., 27 Oct 2025, Diaz, 2022).

1. Classical scalar formulation

In the classical entire-space setting, the Keller–Osserman theorem concerns

$\Delta u=f(u)\quad \text{in }\mathbb{R}^n,$

with $f:[0,\infty)\to[0,\infty)$ continuous and nondecreasing. A nontrivial solution or subsolution exists if and only if

$$\int_{1}^{\infty}\frac{dt}{\sqrt{F(t)}}=\infty, \qquad F(t)=\int_{0}^{t}f(s)\,ds.$$

This integral criterion links the growth of $f$ at infinity to global solvability on $\mathbb{R}^n$ (Sirakov et al., 27 Oct 2025).

The same literature also uses an associated integral map, rather than only the divergence test itself, as the operative “transform.” For large-solution problems with absorption term $B(u)$, one introduces

$F(t)=\int_0^t f(s)\,ds$0

so that $F(t)=\int_0^t f(s)\,ds$1 becomes the canonical blow-up profile near the boundary. This usage shows that the “condition” and the “transform” are closely related but not identical: the first is an integrability threshold, whereas the second is the integral change of variables or its inverse that makes that threshold effective in estimates and asymptotics (Diaz, 2022).

2. Boundary blow-up and inverse profile constructions

For boundary blow-up problems, the Keller–Osserman transform is often an inverse profile rather than an existence criterion for entire-space solutions. In semilinear equations of the form

$F(t)=\int_0^t f(s)\,ds$2

with uniform ellipticity near the boundary, increasing $F(t)=\int_0^t f(s)\,ds$3, and $F(t)=\int_0^t f(s)\,ds$4, the Keller–Osserman integral

$F(t)=\int_0^t f(s)\,ds$5

governs the existence of large solutions, while the inverse transform $F(t)=\int_0^t f(s)\,ds$6 governs their boundary rate. Under the stated assumptions, large solutions satisfy

$F(t)=\int_0^t f(s)\,ds$7

for some constant $F(t)=\int_0^t f(s)\,ds$8 determined by the ellipticity and the data; when $F(t)=\int_0^t f(s)\,ds$9 is increasing, the same boundary profile underlies uniqueness, since two large solutions satisfy $\int_1^\infty \frac{dt}{\sqrt{F(t)}}$0 (Diaz, 2022).

For the $\int_1^\infty \frac{dt}{\sqrt{F(t)}}$1-Hessian operator, the transform is modified to reflect the fully nonlinear structure. If

$\int_1^\infty \frac{dt}{\sqrt{F(t)}}$2

then the boundary blow-up problem

$\int_1^\infty \frac{dt}{\sqrt{F(t)}}$3

admits a positive blow-up solution $\int_1^\infty \frac{dt}{\sqrt{F(t)}}$4 if and only if

$\int_1^\infty \frac{dt}{\sqrt{F(t)}}$5

The paper writes the corresponding Keller–Osserman transform as

$\int_1^\infty \frac{dt}{\sqrt{F(t)}}$6

When $\int_1^\infty \frac{dt}{\sqrt{F(t)}}$7, this reduces to the classical Laplacian case; for $\int_1^\infty \frac{dt}{\sqrt{F(t)}}$8, it reaches the Monge–Ampère regime. The proof uses comparison, sub- and supersolutions, and radial reduction in balls, while the ambient domain is required to be strictly $\int_1^\infty \frac{dt}{\sqrt{F(t)}}$9-convex with smooth boundary (Covei, 2015).

3. Coupled systems and weighted radial problems

In elliptic systems, the Keller–Osserman transform ceases to be a single scalar formula and instead encodes coupling. For the quasilinear system

$K^{-1}$0

the relevant system condition is

$K^{-1}$1

Under the stated monotonicity assumptions on $K^{-1}$2, this yields infinitely many positive entire radial solutions, while additional integral conditions involving $K^{-1}$3 and $K^{-1}$4 separate bounded solutions from large solutions $K^{-1}$5 as $K^{-1}$6 (Covei, 2011).

Weighted semilinear systems use analogous transforms componentwise. One formulation introduces

$K^{-1}$7

with the existence criterion

$K^{-1}$8

Another formulation writes

$K^{-1}$9

and then uses $\Delta u=f(u)\quad \text{in }\mathbb{R}^n,$0 to derive pointwise bounds and to distinguish bounded from unbounded radial entire solutions under weight-tail conditions (Covei, 2015, Covei, 2016).

More recent system papers make the transform explicitly reciprocal. For

$\Delta u=f(u)\quad \text{in }\mathbb{R}^n,$1

one introduces

$\Delta u=f(u)\quad \text{in }\mathbb{R}^n,$2

under the finite integral conditions

$\Delta u=f(u)\quad \text{in }\mathbb{R}^n,$3

These transforms function as subharmonic Lyapunov-type quantities, yielding existence of infinitely many positive entire radial solutions for a nonempty set of central values, closedness of the admissible set of central data, and largeness at boundary points of that set (Covei, 4 Sep 2025).

A competitive-system variant reduces the pair $\Delta u=f(u)\quad \text{in }\mathbb{R}^n,$4 to the scalar sum $\Delta u=f(u)\quad \text{in }\mathbb{R}^n,$5 and applies

$\Delta u=f(u)\quad \text{in }\mathbb{R}^n,$6

or, under a monotone lower envelope $\Delta u=f(u)\quad \text{in }\mathbb{R}^n,$7,

$\Delta u=f(u)\quad \text{in }\mathbb{R}^n,$8

The existence theory is then controlled by the finite generalized Keller–Osserman integral

$\Delta u=f(u)\quad \text{in }\mathbb{R}^n,$9

The proof is organized around monotone iteration, reduction to a scalar inequality for $f:[0,\infty)\to[0,\infty)$0, application of the transform, and a two-step radial integration argument (Covei, 15 Sep 2025).

Even where comparison principles are absent, Keller–Osserman-type mechanisms persist as a priori estimates. For quasilinear absorption or mixed systems, if

$f:[0,\infty)\to[0,\infty)$1

then nonnegative weak solutions satisfy

$f:[0,\infty)\to[0,\infty)$2

These are described as Keller–Osserman type estimates, and for mixed systems one component satisfies a Harnack inequality (Bidaut-Véron et al., 2011).

4. Gradient dependence, geometric settings, and rough coefficients

For quasilinear inequalities with gradient terms, the transform becomes operator-adapted. On the Heisenberg group, for

$f:[0,\infty)\to[0,\infty)$3

one defines

$f:[0,\infty)\to[0,\infty)$4

and the generalized Keller–Osserman condition is

$f:[0,\infty)\to[0,\infty)$5

For the inequality

$f:[0,\infty)\to[0,\infty)$6

the transform is modified through

$f:[0,\infty)\to[0,\infty)$7

and the relevant condition becomes

$f:[0,\infty)\to[0,\infty)$8

In the $f:[0,\infty)\to[0,\infty)$9-Laplacian case, these conditions recover the classical threshold and are proved sharp; the same framework also extends, with minor modifications, to Euclidean space (Magliaro et al., 2010).

On Riemannian manifolds, the transform is intertwined with geometry. For coercive inequalities of the form

$$\int_{1}^{\infty}\frac{dt}{\sqrt{F(t)}}=\infty, \qquad F(t)=\int_{0}^{t}f(s)\,ds.$$0

one introduces

$$\int_{1}^{\infty}\frac{dt}{\sqrt{F(t)}}=\infty, \qquad F(t)=\int_{0}^{t}f(s)\,ds.$$1

and then the generalized Keller–Osserman conditions are

$$\int_{1}^{\infty}\frac{dt}{\sqrt{F(t)}}=\infty, \qquad F(t)=\int_{0}^{t}f(s)\,ds.$$2

These thresholds interact with curvature bounds, volume growth, weak and strong maximum principles at infinity, compact support principles, and Liouville-type properties. The paper emphasizes that necessity is often more geometry-independent than sufficiency, and introduces “fake distance functions” via nonlinear Green kernels to carry the method beyond manifolds with a pole (Bianchini et al., 2018).

For divergence-form elliptic operators with unbounded lower-order coefficients,

$$\int_{1}^{\infty}\frac{dt}{\sqrt{F(t)}}=\infty, \qquad F(t)=\int_{0}^{t}f(s)\,ds.$$3

explicit radial barriers are no longer available. In that setting, the Keller–Osserman transform is described as an ODE for the measure of level sets,

$$\int_{1}^{\infty}\frac{dt}{\sqrt{F(t)}}=\infty, \qquad F(t)=\int_{0}^{t}f(s)\,ds.$$4

This replaces the classical radial transform and yields a generalized Harnack inequality, the optimal strong maximum principle condition

$$\int_{1}^{\infty}\frac{dt}{\sqrt{F(t)}}=\infty, \qquad F(t)=\int_{0}^{t}f(s)\,ds.$$5

and the extension of the classical entire-space Keller–Osserman theorem to operators with locally unbounded coefficients in uniformly local Lebesgue spaces (Sirakov et al., 27 Oct 2025).

5. Higher-order, Hessian, and fractional generalizations

The transform also survives changes in differential order. For higher-order differential inequalities

$$\int_{1}^{\infty}\frac{dt}{\sqrt{F(t)}}=\infty, \qquad F(t)=\int_{0}^{t}f(s)\,ds.$$6

with $$\int_{1}^{\infty}\frac{dt}{\sqrt{F(t)}}=\infty, \qquad F(t)=\int_{0}^{t}f(s)\,ds.$$7 nondecreasing and convex, the generalized Keller–Osserman condition is

$$\int_{1}^{\infty}\frac{dt}{\sqrt{F(t)}}=\infty, \qquad F(t)=\int_{0}^{t}f(s)\,ds.$$8

If, in addition,

$$\int_{1}^{\infty}\frac{dt}{\sqrt{F(t)}}=\infty, \qquad F(t)=\int_{0}^{t}f(s)\,ds.$$9

then every global weak solution is trivial; if

$f$0

then there are no nontrivial global weak solutions at all. For $f$1, the criterion reduces to the classical second-order form (Kon'kov et al., 2018).

In the fractional Laplacian problem

$f$2

the transform reappears in the study of “very large solutions,” namely $f$3 solutions with boundary singularity stronger than $f$4. The paper defines

$f$5

and proves existence under the sufficient condition

$f$6

An additional hypothesis,

$f$7

is shown to be equivalent to

$f$8

The resulting solutions satisfy

$f$9

For $\mathbb{R}^n$0, very large solutions exist if and only if

$\mathbb{R}^n$1

and then

$\mathbb{R}^n$2

This is presented as a fractional Keller–Osserman condition (Abatangelo, 2014).

6. Conceptual scope and recurrent ambiguities

The cited literature shows that the expression “Keller–Osserman transform” does not refer to one canonical formula. It denotes, depending on the problem class, an inverse blow-up profile such as $\mathbb{R}^n$3, an operator-dependent integral map $\mathbb{R}^n$4 used through $\mathbb{R}^n$5, a scalarization of a coupled system through transforms like $\mathbb{R}^n$6, $\mathbb{R}^n$7, or $\mathbb{R}^n$8, or a level-set ODE replacing explicit radial reductions (Diaz, 2022, Magliaro et al., 2010, Sirakov et al., 27 Oct 2025).

A recurring source of confusion is the sign of the decisive integral condition. For entire solvability of $\mathbb{R}^n$9 on $B(u)$0, the threshold is the divergence of

$B(u)$1

for boundary blow-up problems, the corresponding large-solution criterion is typically the finiteness of an analogous integral, as in the semilinear and $B(u)$2-Hessian settings. The same family of transforms therefore governs qualitatively different outcomes—existence of entire solutions, existence of boundary blow-up solutions, Liouville triviality, strong maximum principles, Harnack inequalities, and sharp a priori estimates—because the transform is tied to the specific operator, geometry, and direction of the comparison argument rather than to a single universal theorem (Sirakov et al., 27 Oct 2025, Covei, 2015).

A second ambiguity concerns whether the transform is merely a formal change of variables. The available results indicate a broader role. In some settings it produces explicit asymptotic rates; in others it identifies the admissible set of central values for radial systems, yields subharmonic or Lyapunov-type functionals, or substitutes for barrier constructions that fail in the presence of unbounded coefficients or non-Euclidean geometry. This suggests that the Keller–Osserman transform is best understood as the integral analytic core of a method, with the precise formula determined by the nonlinear operator and the qualitative property under study (Covei, 4 Sep 2025, Covei, 15 Sep 2025, Bianchini et al., 2018).