Truncation-Penalization of the Critical Term

• Truncation-penalization is a method that modifies critical nonlinear terms using smooth cutoff functions to suppress concentration and restore compactness in variational frameworks.
• It constructs auxiliary penalized energy functionals that match the original system below a set threshold while truncating problematic terms beyond to prevent blowup.
• These techniques enable rigorous existence, multiplicity, and asymptotic analysis in fractional, nonlocal, and stochastic models by effectively controlling singular behaviors.

Truncation-penalization of a critical term refers to a class of analytical and variational techniques designed to control and manage the effects of singular, concentration-prone, or non-compact nonlinearities—typically at critical exponent—within functional-analytic or PDE frameworks. Foundational examples come from critical elliptic and nonlocal problems (such as fractional Choquard and Hardy–Sobolev equations), as well as from stochastic processes with critical scaling properties, where truncating or penalizing problematic terms allows for the recovery of compactness or the derivation of sharp asymptotic behaviors. This entry summarizes key methodologies, implementations, and theoretical implications of truncation-penalization devices as detailed in recent research.

1. Rationale and Variational Setting

Critical nonlinearities are notorious for undermining compactness—i.e., Palais–Smale sequences may fail to converge due to concentration phenomena or scale-invariance. Truncation-penalization modifies either the nonlinearity or the energy landscape to suppress or eliminate such escape routes while maintaining correspondence with the original problem on regions of interest.

For the upper-critical fractional Choquard equation, the variational framework targets normalized solutions on the $L^2$-sphere: $$S(a) = \{ u \in H^s(\mathbb{R}^N) : \|u\|_{L^2}^2 = a \}$$ with the energy functional

$$E(u) = \frac{1}{2} \int_{\mathbb{R}^N} |(-\Delta)^{s/2}u|^2 + V(\varepsilon x) |u|^2 \, dx - \frac{1}{2p} \int_{\mathbb{R}^N} (I_\alpha * |u|^p)|u|^p \, dx - \frac{1}{2q} \int_{\mathbb{R}^N} (I_\alpha * |u|^q) |u|^q \, dx$$

where the Hartree (critical) term—typically at the upper-end exponent $p = (N+\alpha)/(N-2s)$—admits $L^2$-preserving dilations, obstructing compactness and necessitating further intervention (Aikyn et al., 30 Nov 2025).

2. Formulation of Truncation-Penalization Devices

Truncation-penalization schemes employ auxiliary cutoff functions and truncation radii, with smooth transition from “no penalty” in the regime of interest to “full suppression” beyond a prescribed norm threshold.

For instance, let $0 < R_0 < R_1$ be fixed radii and choose a smooth, non-increasing cutoff $\tau \in C^\infty([0,\infty); [0,1])$ with $\tau(r) = 1$ for $r \le R_0$, $\tau(r) = 0$ for $r \ge R_1$. Define the penalized energy: $$J_{\varepsilon,T}(u) = \frac{1}{2} \|(-\Delta)^{s/2} u\|_2^2 + \frac{1}{2} \int V(\varepsilon x)|u|^2 dx - \frac{\tau(\|u\|_{H^s})}{2p} \int (I_\alpha * |u|^p)|u|^p - \frac{1}{2q}\int (I_\alpha*|u|^q)|u|^q$$ or

$$J_{\varepsilon,T}(u) = J_\varepsilon(u) - P(u), \qquad P(u) = \frac{1-\tau(\|u\|_{H^s})}{2p}\int (I_\alpha * |u|^p) |u|^p.$$

The penalty $P(u)$ is active only when $\|u\|_{H^s} \ge R_0$, becoming maximal (full truncation of the critical term) for $\|u\|_{H^s} \ge R_1$ (Aikyn et al., 30 Nov 2025).

This results in an energy functional that coincides with the original for $\|u\|_{H^s} < R_0$, but which blocks supercritical blowup and concentration for high norm regimes.

3. Compactness Restoration and Critical-Point Theory

Palais–Smale compactness can be restored below a certain energy threshold. For the penalized functional $J_{\varepsilon,T}$:

• Any (PS)$_c$ sequence with $c < b_{\infty,T,a}$ is bounded in $H^s$ (Lemma 3.1).
• Vanishing of mass is excluded (Lemmas 3.2-3.3), exploiting sharp interpolation inequalities and splitting lemmas.
• Once $\|u\|_{H^s} \ge R_1$, the critical term is zero, hence sequences cannot concentrate via critical nonlocal nonlinearities.

Variational methods (mountain-pass, fibering maps) yield min–max levels that ensure solutions exist within domains where the truncation is inactive, provided the $L^2$-norm (mass parameter $a$) is sufficiently small: $$a \le \big[K_q^{-1} S_\alpha^{\vartheta_q}\big]^{\frac{1}{q(1-\gamma_{q,s})}}$$ guarantees that all relevant maximizers $u_{t(u)}$ remain within the ball $\|u\|_{H^s} < R_0$, so $P(u) = 0$ and criticality is not penalized (Aikyn et al., 30 Nov 2025).

4. Interplay with Concentration-Compactness and Penalization in Related Models

Truncation-penalization extends and refines classical penalization methods used for supercritical Schrödinger-type problems. For example:

• In fractional Schrödinger settings with mixed subcritical/supercritical nonlinearities, penalization is staged (first spatially “outside the well”; next on the supercritical term at large heights). Truncations are shown not to alter solutions of interest by careful a priori $L^\infty$ estimates (via Moser iterations and extensions), ensuring that true solutions never reach the truncation “danger zone” (Ambrosio, 2017).
• In fractional Hardy–Sobolev frameworks, explicit quantitative truncation estimates for ground states yield a penalization mechanism wherein the “loss” produced by truncating the critical term outweighs the accompanying “loss” in nonlocal seminorm, thus raising the energy of any would-be concentrating sequence and preventing concentration at the origin (Marano et al., 2019).

This methodology is fundamental for the Brezis–Nirenberg type analysis in nonlocal or critical exponents, establishing existence, multiplicity, and concentration of solutions.

5. Broader Context: Truncated-Penalization in Stochastic and Functional Analysis

Analogues of truncation-penalization appear in stochastic processes with heavy-tailed or critical scaling. For perpetuities at criticality ($E[\log A]=0$), classical tail analyses (Kesten's theorem) are supplanted by the paper of truncated κ-th moments: $M(t) := E[U^\kappa 1_{U \le t}]$ which satisfy asymptotic expansions under minimal regularity assumptions. The penalization aspect enters by considering the increment

$$R(t) := E[(A U + B)^\kappa 1_{\{A U+B \le t\}}] - E[(A U)^\kappa 1_{\{A U \le t\}}]$$

enabling quantification of the “critical” contribution without requiring global moment regularity, and yielding new limit theorems (LLN and CLT) for GARCH processes and other critical recursions (Jakubowski et al., 2020).

6. Summary Table: Key Features of Truncation-Penalization Devices

Model/Setting	Truncation Mechanism	Main Role
Fractional Choquard (Aikyn et al., 30 Nov 2025)	Smooth norm cutoff τ on upper-critical nonlinearity	Prevent critical blowup, restore (PS) compactness
Fractional Schrödinger (Ambrosio, 2017)	Subcritical+supercritical truncation (in $x$, $t$)	Localize solutions, control over supercritical growth
Fractional Hardy-Sobolev (Marano et al., 2019)	Linear cut-off on rescaled Aubin–Talenti bubble	Penalize concentration, force energy separation
Perpetuity and SRE models (Jakubowski et al., 2020)	Truncated moment functionals $M(t)$	Bypass tail regularity, derive sharp asymptotics

7. Theoretical and Practical Implications

Truncation-penalization of the critical term is now standard in the analysis of PDEs and stochastic processes at or beyond criticality. Its principal impact is:

• Enabling rigorous existence and multiplicity results in upper-critical or $L^2$-supercritical regimes, where direct variational or probabilistic methods typically fail due to lack of compactness or integrability (Aikyn et al., 30 Nov 2025, Ambrosio, 2017).
• Yielding a robust method to circumvent concentration, compactness loss, or infinite expectation phenomena without imposing strict regularity or decay assumptions on the nonlinearity or process coefficients (Jakubowski et al., 2020).
• Providing explicit, quantitative penalization estimates and energy gaps, essential for the fine analysis of concentration, blow-up, and asymptotic variational geometry (Marano et al., 2019).

A plausible implication is that such devices can be tailored to increasingly general nonlocal, critical, or stochastic models, provided the truncation preserves the structure required on the solution manifold of interest.

---

References:

(Aikyn et al., 30 Nov 2025, Ambrosio, 2017, Marano et al., 2019, Jakubowski et al., 2020)