---
title: Second-Order Limiting Subdifferential
url: https://www.emergentmind.com/topics/second-order-limiting-subdifferential
type: topic
---

# Second-Order Limiting Subdifferential

A second-order limiting subdifferential is the central generalized second-order object in nonsmooth variational analysis, serving as the canonical extension of the classical Hessian for lower semicontinuous, nonconvex, and nonsmooth functions. Originating with the works of Mordukhovich and Rockafellar, it underpins much of contemporary nonsmooth optimization, variational stability, and generalized differentiation theory. This construction enables precise second-order optimality conditions, comprehensive calculus rules, and characterizations of criticality, stability, and metric regularity in a broad class of nonsmooth problems [2312.16277, 1110.4572, 1304.7385].

## 1. Definition and Construction

Let $f:\mathbb{R}^n \to \mathbb{R}\cup\{+\infty\}$ be a proper lower semicontinuous function, prox-regular at $\bar x$ for $\bar v \in \partial f(\bar x)$. The first-order (Mordukhovich/limiting) subdifferential is
\[
\partial f(x) := \limsup_{(x',v')\to(x,v),\, v'\in\hat\partial f(x')} \{ v'\},
\]
where $\hat\partial f$ denotes the Fréchet subdifferential. The graph of $\partial f$ is
\[
\mathrm{gph}\,\partial f = \{\, (x,v): v\in\partial f(x)\, \}.
\]
The **second-order limiting subdifferential** (generalized Hessian) of $f$ at $(\bar x,\bar v)$ in direction $w\in\mathbb{R}^n$ is defined [2312.16277, 1110.4572, 1304.7385] as
\[
\partial^2 f(\bar x\mid \bar v)(w) := D^*(\partial f)(\bar x,\bar v)(w) = \{\, z\in\mathbb{R}^n : (z,-w)\in N_{\mathrm{gph}\,\partial f}(\bar x,\bar v) \,\},
\]
where $D^*$ denotes the Mordukhovich coderivative and $N_{\mathrm{gph}\,\partial f}$ is the limiting normal cone to the graph.

In the $C^2$-smooth case, $\partial^2 f(\bar x\mid \nabla f(\bar x))(w) = \{\nabla^2 f(\bar x)w\}$, recovering the classical Hessian [2312.16277, 1304.7385].

## 2. Key Properties

The second-order limiting subdifferential possesses several structural and analytical properties [2312.16277, 1110.4572, 2503.00927, 1909.09795]:

- **Closed Graph**: $\mathrm{gph}\,\partial^2 f(\bar x\mid \bar v)$ is closed and conic in $\mathbb{R}^n\times\mathbb{R}^n$.
- **Homogeneity**: $\lambda\geq0$ implies $\partial^2 f(\bar x\mid \bar v)(\lambda w) = \lambda \partial^2 f(\bar x\mid \bar v)(w)$.
- **Local Boundedness**: The mapping $w \mapsto \partial^2f(\bar x\mid \bar v)(w)$ is locally bounded and upper semicontinuous.
- **Positive (Semi)definiteness**: At a local minimizer $\bar x$ with $0\in\partial f(\bar x)$, every $z\in\partial^2 f(\bar x \mid 0)(w)$ satisfies $(z,w) \geq 0$ for all $w$; strict definiteness characterizes tilt stability.
- **Reduction to Hessian**: When $f$ is $C^2$ at $\bar x$, $\partial^2 f(\bar x\mid \nabla f(\bar x))(w) = \{ \nabla^2 f(\bar x)w \}$.
- **Nonemptiness and Compactness**: For $C^{1,1}$ functions, $\partial^2 f(\bar x\mid \nabla f(\bar x))(w)$ is a nonempty compact subset of $\mathbb{R}^n$ [2503.00927].

## 3. Calculus Rules

Second-order limiting subdifferentials allow robust calculus, supporting sum, chain, and scalar multiplication rules under suitable regularity conditions [2312.16277, 1110.4572, 1909.09795, 1710.07469]:

| Operation     | Formula (under qualification conditions)                 | Context                        |
|---------------|---------------------------------------------------------|--------------------------------|
| Sum           | $\partial^2(f+g)(\bar x\mid v)(w) \subset \partial^2 f(\bar x\mid v_f)(w) + \partial^2 g(\bar x\mid v_g)(w)$ | $f, g$ prox-regular at $\bar x$, $v = v_f+v_g$ |
| Chain         | $\partial^2 (h\circ F)(\bar x\mid F'(\bar x)^T v_h)(w) \subset F'(\bar x)^T \partial^2 h(\bar y\mid v_h)(F'(\bar x)w) + \nabla^2 F(\bar x)(w,\zeta)$ | $F$ $C^2$, $h$ prox-regular |
| Scalar Mult.  | $\partial^2(\alpha f)(\bar x\mid \alpha \bar v) = \alpha \partial^2 f(\bar x\mid \bar v)$ | $\alpha>0$                    |

Equality in the sum and chain rules requires stronger conditions (e.g., full rank, graphical regularity) [1110.4572, 1304.7385].

## 4. Second-Order Optimality and Stability

The second-order limiting subdifferential characterizes both necessary and sufficient optimality conditions in nonsmooth and possibly nonconvex problems [2312.16277, 1304.7385, 1204.5794]:

- **Necessary Condition**: If $\bar x$ is a local minimizer and $0\in \partial f(\bar x)$,
  \[
  (z,w)\geq 0 \qquad \forall\, z\in\partial^2 f(\bar x\mid 0)(w),\ \forall\,w\in\mathbb{R}^n.
  \]
- **Sufficient Condition**: If $\exists\,\rho>0$ s.t.\ $(z,w)\geq \rho\|w\|^2$ for all nearby $(x,v)$, $z\in\partial^2 f(x\mid v)(w)$, then $\bar x$ is a strict local minimizer.
- **Tilt Stability**: If $f$ is subdifferentially continuous and there is $\rho>0$ so that
  \[
  (z,w) > \rho\|w\|^2 \quad \forall w\neq 0,\, z\in\partial^2 f(\bar x\mid 0)(w),
  \]
  then $\bar x$ is tilt-stable with modulus $1/\rho$, i.e., the argmin mapping $y\mapsto \operatorname{argmin}_{x\approx \bar x}\{f(x)-\langle y,x-\bar x\rangle\}$ is single-valued and Lipschitz [2312.16277, 1204.5794, 1304.7385].

## 5. Explicit Computations and Classical Cases

For convex piecewise-linear (PL) functions and $C^2$ functions, the second-order subdifferential admits explicit expressions [2312.16277, 1110.4572, 1609.07199]:

- **Convex PL**: $f(x) = \max_{i} \{ a_i^T x + b_i \}$ yields $\partial f(\bar x) = \mathrm{conv}\{ a_i : i\in I(\bar x)\}$ and
  \[
  \partial^2 f(\bar x\mid \bar v)(w) = \{0\}
  \]
  for all $w$, reflecting the absence of curvature.
- **$C^2$ Reduction**: For $f$ $C^2$ at $\bar x$,
  \[
  \partial^2f(\bar x\mid \nabla f(\bar x))(w) = \{ \nabla^2 f(\bar x)w \}.
  \]
- **PLQ Functions and Compositions**: For fully amenable composite functions $f(x)=\theta(h(x))$ with $h$ $C^2$, $\theta$ convex PLQ, the chain rule applies and computable formulas for $\partial^2 f$ and critical cones are available [1110.4572, 1609.07199].

## 6. Applications: Constrained Optimization, Critical Multipliers, and Stability

The second-order limiting subdifferential underpins advanced topics in optimization, including KKT multipliers, error bounds, stability, and variational inclusions [2503.00927, 1609.07199, 1110.4572]:

- **KKT Systems**: Second-order necessary and sufficient KKT-type conditions for $C^{1,1}$ vector optimization with constraints require the presence of $\partial^2$ in multiplier and critical cone expressions [2503.00927].
- **Critical Multipliers**: The criticality of Lagrange multipliers in variational/KKT systems is characterized via the second-order limiting subdifferential, with explicit conic-algebraic criteria for piecewise-linear constraints [1609.07199].
- **Strong Regularity, Quadratic Growth, and Tilt Stability**: Weak and strong metric regularity, uniform quadratic growth, and tilt stability are all characterized via positive definiteness of the second-order (limiting) subdifferential [1204.5794, 1304.7385]. The equivalences enable robust stability and error-bound guarantees in nonconvex and nonsmooth settings.

## 7. Relations to Other Second-Order Constructions

The second-order limiting subdifferential is closely related to, but generally finer than, other second-order objects:

- **Clarke’s Generalized Hessian**: In $C^{1,1}$ settings, the Mordukhovich construction refines Clarke’s generalized Jacobian by capturing additional directional information ("hidden curvature") provided by the normal cone to the graph [1909.09795, 2503.00927].
- **Second Subderivative**: The variational second subderivative $d^2 f(\bar x|\bar v)(u)$ relates closely to the coderivative-based $\partial^2 f(\bar x\mid \bar v)$ and provides variational characterizations of tilt stability and quadratic growth [1204.5794, 1304.7385].
- **Operator Inclusions and Abstract Nonsmooth Calculus**: The second-order limiting subdifferential extends to Banach spaces and set-valued mappings, supporting extremal conditions in operator-inclusion problems [1710.07469].

---

**References**:  
- "Second-Order Subdifferential Optimality Conditions in Nonsmooth Optimization" [2312.16277]  
- "Second-order subdifferential calculus with applications to tilt stability in optimization" [1110.4572]  
- "Second-order growth, tilt stability, and metric regularity of the subdifferential" [1304.7385]  
- "On second-order Karush--Kuhn--Tucker optimality conditions for $C^{1,1}$ vector optimization problems" [2503.00927]  
- "Second-order optimality conditions for multiobjective optimization problems with constraints" [1909.09795]  
- "Tilt stability, uniform quadratic growth, and strong metric regularity of the subdifferential" [1204.5794]  
- "Critical Multipliers in Variational Systems via Second-order Generalized Differentiation" [1609.07199]  
- "Second-order subdifferential. Extremal problems for operational inclusions" [1710.07469]

Source: https://www.emergentmind.com/topics/second-order-limiting-subdifferential