Kinematic Tangent Cone

• Kinematic tangent cone is defined by realizable motions—feasible arcs and rescaling—that reveal the true infinitesimal structure without relying solely on linear approximations.
• It plays a crucial role in singular optimization and generalized equations by addressing degenerate cases where traditional derivatives fail.
• It underpins studies in geometric measure theory and complex geometry, linking concepts such as paratangent cones, blow-up techniques, and strict differentiability.

The kinematic tangent cone is a tangent-cone notion defined through actual infinitesimal motions rather than solely through first-order algebraic linearization. In the most explicit use of the term, for singular generalized equations the cone consists of directions $h$ for which the solution set admits feasible arcs of the form $x(t)=x_0+th+o(t)$ (Prusinska et al., 2018). Closely related Euclidean formulations replace a fixed base point by pairs of nearby points, leading to paratangent cones that encode infinitesimal chordal motion and strict differentiability (Bigolin et al., 2012). In a broader geometric-measure-theoretic usage, tangent cones obtained by blow-up under rescaling are likewise described as arising from a kinematic dilation process, although not always under the literal name “kinematic tangent cone” (Bellettini, 2011).

1. Terminological scope and defining intuition

The expression “kinematic tangent cone” is used most directly for solution sets of singular inclusions, where tangency is defined by the existence of nearby feasible motions rather than by the ordinary derivative (Prusinska et al., 2018). In that setting, the central object is the solution set

$M_{F(x_0)}:=\{x\in X:\ 0\in f(x)+N_C(x)\},$

associated with a generalized equation

$0\in F(x):=f(x)+N_C(x).$

A vector $h$ belongs to the tangent cone when the constraint admits a path $x_0+th+o(t)$ remaining feasible for small $t$ (Prusinska et al., 2018).

This usage is consistent with an older geometric distinction between tangent and paratangent behavior. For a set $F\subset \mathbb{R}^n$, the upper tangent cone records infinitesimal directions from a fixed base point, while the upper paratangent cone allows the base point to move and is therefore the “kinematic” refinement (Bigolin et al., 2012). The same motif appears in other areas: some authors describe tangent cones produced by blow-up maps

$T_{x_0,r}=\left(\frac{x-x_0}{r}\right)_*T$

as arising from a kinematic blow-up or dilation process, even when no separate named object is introduced (Bellettini, 2011).

A plausible implication is that “kinematic tangent cone” is best understood not as a single universal formalism, but as a family of tangent constructions unified by one principle: infinitesimal geometry is extracted from realizable motions, secants, or rescalings rather than from a purely static linear approximation.

2. Kinematic definition for singular generalized equations

In the formulation of singular inclusions, the generalized equation is

$0 \in f(x,y)+Q(y),$

and in the main singular setting,

$x(t)=x_0+th+o(t)$0

where $x(t)=x_0+th+o(t)$1 is smooth, $x(t)=x_0+th+o(t)$2 is a convex set, and

$x(t)=x_0+th+o(t)$3

The distinguishing feature is that surjectivity of the derivative is not assumed; the analysis is explicitly directed at the singular or degenerate case (Prusinska et al., 2018).

The kinematic tangent cone to the solution set is defined as follows. A vector $x(t)=x_0+th+o(t)$4 belongs to $x(t)=x_0+th+o(t)$5 if for every sufficiently small $x(t)=x_0+th+o(t)$6 there exists a correction $x(t)=x_0+th+o(t)$7 such that

$x(t)=x_0+th+o(t)$8

and

$x(t)=x_0+th+o(t)$9

Equivalently, $M_{F(x_0)}:=\{x\in X:\ 0\in f(x)+N_C(x)\},$0 is tangent precisely when the solution set admits a feasible arc

$M_{F(x_0)}:=\{x\in X:\ 0\in f(x)+N_C(x)\},$1

This is the exact reason the paper calls the object a kinematic tangent cone: membership is certified by an actual nearby feasible motion (Prusinska et al., 2018).

The construction is especially significant in the completely degenerate regime. The paper assumes

$M_{F(x_0)}:=\{x\in X:\ 0\in f(x)+N_C(x)\},$2

and

$M_{F(x_0)}:=\{x\in X:\ 0\in f(x)+N_C(x)\},$3

for some $M_{F(x_0)}:=\{x\in X:\ 0\in f(x)+N_C(x)\},$4. Thus the ordinary first derivative does not control local geometry; the first nonzero term occurs at order $M_{F(x_0)}:=\{x\in X:\ 0\in f(x)+N_C(x)\},$5 (Prusinska et al., 2018). This suggests that the cone is genuinely higher order: it is shaped by feasible arcs compatible with the $M_{F(x_0)}:=\{x\in X:\ 0\in f(x)+N_C(x)\},$6-th order expansion and the normal-cone geometry.

3. Higher-order tangent cone theorem and $M_{F(x_0)}:=\{x\in X:\ 0\in f(x)+N_C(x)\},$7-regularity

The main tangent-cone result for singular inclusions is presented as a generalized Lusternik theorem. Assume

• $M_{F(x_0)}:=\{x\in X:\ 0\in f(x)+N_C(x)\},$8,
• $M_{F(x_0)}:=\{x\in X:\ 0\in f(x)+N_C(x)\},$9 for $0\in F(x):=f(x)+N_C(x).$0,
• $0\in F(x):=f(x)+N_C(x).$1,
• and $0\in F(x):=f(x)+N_C(x).$2 satisfies the $0\in F(x):=f(x)+N_C(x).$3-kernel condition

$0\in F(x):=f(x)+N_C(x).$4

meaning

$0\in F(x):=f(x)+N_C(x).$5

If the associated multifunction also satisfies the strong $0\in F(x):=f(x)+N_C(x).$6-regularity estimate along $0\in F(x):=f(x)+N_C(x).$7,

$0\in F(x):=f(x)+N_C(x).$8

then

$0\in F(x):=f(x)+N_C(x).$9

This is Theorem 3 in the paper (Prusinska et al., 2018).

The result shifts the governing infinitesimal object from the ordinary derivative to the first nonvanishing higher derivative. In concise form, the theorem asserts that

$h$0

and the strong $h$1-regularity condition holds along $h$2 (Prusinska et al., 2018).

The same higher-order mechanism underlies the singular implicit function theorem in the paper. For $h$3 near $h$4, there exists a local implicit solution $h$5 such that

$h$6

with estimate

$h$7

The key assumptions are the Banach condition, strong $h$8-regularity along $h$9, and the $x_0+th+o(t)$0-factor approximation condition (Prusinska et al., 2018). In this framework, the kinematic tangent cone is not an isolated geometric definition; it is the tangent-level counterpart of a higher-order implicit-function theory for degenerate generalized equations.

4. Paratangent cones as the Euclidean kinematic model

A closely related Euclidean theory formulates the kinematic content in terms of paratangent cones. For a subset $x_0+th+o(t)$1 and $x_0+th+o(t)$2, the paper considers four cones: $x_0+th+o(t)$3 The upper tangent cone is based on sequences from the fixed base point, whereas the upper paratangent cone is defined by allowing both endpoints of secants to vary toward $x_0+th+o(t)$4 (Bigolin et al., 2012).

The upper paratangent cone has the sequential characterization

$x_0+th+o(t)$5

This is the “kinematic” version because it captures infinitesimal relative motions of nearby points rather than directions based at a single point (Bigolin et al., 2012). The cones satisfy

$x_0+th+o(t)$6

The geometric force of this refinement is expressed by the Four-cones Coincidence Theorem: a non-empty subset $x_0+th+o(t)$7 is a $x_0+th+o(t)$8-manifold if and only if $x_0+th+o(t)$9 is locally compact and

$t$0

When the extreme paratangent cones coincide, all four cones coincide at every point (Bigolin et al., 2012).

The same paper makes the link to strict differentiability explicit. A function $t$1 is strictly differentiable at $t$2 if there exists linear $t$3 such that

$t$4

and this is equivalent to

$t$5

Accordingly, the upper paratangent cone is the geometric incarnation of strict differentiability (Bigolin et al., 2012). This suggests that, in Euclidean geometry, the kinematic tangent cone is most naturally realized by the paratangent cone rather than by the ordinary contingent cone.

A distinct but compatible formulation appears for definable sets in an o-minimal structure. There the tangent cone

$t$6

is paired with the paratangent cone

$t$7

For connected locally closed definable sets, $t$8 at every point is equivalent to $t$9 being a $F\subset \mathbb{R}^n$0 manifold (Kurdyka et al., 2017). Here again the two-point, secant-based cone supplies the kinematic refinement.

5. Blow-up, rescaling, and metric interpretations

In several geometric settings, tangent cones are produced by explicit dilation or blow-up procedures, and the literature summarized here presents this as the closest analogue to a kinematic tangent-cone construction. For positive-$F\subset \mathbb{R}^n$1 De Rham currents in an almost complex manifold, the blow-up at $F\subset \mathbb{R}^n$2 with scale $F\subset \mathbb{R}^n$3 is

$F\subset \mathbb{R}^n$4

and tangent cones are weak limits

$F\subset \mathbb{R}^n$5

along $F\subset \mathbb{R}^n$6 (Bellettini, 2011). The paper states that if “kinematic tangent cone” is understood as the cone obtained by rescaling the current around a point, then this is precisely the notion under study. Under the density-approximation hypothesis

$F\subset \mathbb{R}^n$7

the tangent cone is unique and equals

$F\subset \mathbb{R}^n$8

with $F\subset \mathbb{R}^n$9 a $T_{x_0,r}=\left(\frac{x-x_0}{r}\right)_*T$0-holomorphic disk through the origin (Bellettini, 2011).

For positive plurisubharmonic or plurisuperharmonic currents, the dilation is

$T_{x_0,r}=\left(\frac{x-x_0}{r}\right)_*T$1

and the tangent cone at $T_{x_0,r}=\left(\frac{x-x_0}{r}\right)_*T$2 is the weak limit of $T_{x_0,r}=\left(\frac{x-x_0}{r}\right)_*T$3 as $T_{x_0,r}=\left(\frac{x-x_0}{r}\right)_*T$4, when it exists (Ghiloufi et al., 2011). Sufficient conditions are given in terms of the Lelong functions: $T_{x_0,r}=\left(\frac{x-x_0}{r}\right)_*T$5 in the plurisubharmonic case, and in the plurisuperharmonic case this is supplemented by

$T_{x_0,r}=\left(\frac{x-x_0}{r}\right)_*T$6

Any cluster value is a positive conic pluriharmonic current (Ghiloufi et al., 2011).

A metric version appears for planar starlike sets. For a metric space $T_{x_0,r}=\left(\frac{x-x_0}{r}\right)_*T$7, a point $T_{x_0,r}=\left(\frac{x-x_0}{r}\right)_*T$8, and a normalizing sequence $T_{x_0,r}=\left(\frac{x-x_0}{r}\right)_*T$9, one studies limits of rescaled distance ratios

$0 \in f(x,y)+Q(y),$0

For a starlike set $0 \in f(x,y)+Q(y),$1 with center $0 \in f(x,y)+Q(y),$2, every tangent space $0 \in f(x,y)+Q(y),$3 is isometric to the smallest closed cone with vertex $0 \in f(x,y)+Q(y),$4 containing the set: $0 \in f(x,y)+Q(y),$5 The paper describes this as the precise “kinematic tangent cone” interpretation of the tangent-space construction (Dovgoshey et al., 2012).

These examples do not identify a single common formal definition, but they all realize the same structural idea: infinitesimal geometry is recovered by actual rescaling dynamics.

6. Applications and conceptual significance

In singular optimization and complementarity, the kinematic tangent cone provides a higher-order description of feasible directions when Robinson-type or classical surjectivity assumptions fail. The paper applies its theory to nonlinear complementarity conditions

$0 \in f(x,y)+Q(y),$6

where Theorem 1 applies with $0 \in f(x,y)+Q(y),$7, yielding a local solution mapping and an estimate

$0 \in f(x,y)+Q(y),$8

and to KKT systems for nonlinear programming, where one example gives

$0 \in f(x,y)+Q(y),$9

The tangent-cone message is that feasible directions are governed by the first nonvanishing higher derivative together with normal-cone geometry (Prusinska et al., 2018).

In Euclidean and definable geometry, the kinematic refinement supplied by paratangent cones yields sharp manifold criteria. For locally closed definable sets, $x(t)=x_0+th+o(t)$00 at every point is equivalent to $x(t)=x_0+th+o(t)$01 regularity (Kurdyka et al., 2017); for locally compact subsets of Euclidean space, coincidence of lower and upper paratangent cones characterizes $x(t)=x_0+th+o(t)$02-manifolds (Bigolin et al., 2012). This places the kinematic cone at the center of first-order regularity theory.

In geometric measure theory and complex geometry, blow-up tangent cones control uniqueness and asymptotic structure. For positive-$x(t)=x_0+th+o(t)$03 currents, uniqueness may fail without a no-jump density hypothesis, and the theorem is described as optimal (Bellettini, 2011). For positive plurisubharmonic or plurisuperharmonic currents, explicit growth conditions on $x(t)=x_0+th+o(t)$04 and $x(t)=x_0+th+o(t)$05 guarantee existence of the blow-up limit (Ghiloufi et al., 2011). A plausible implication is that the kinematic perspective is especially powerful in singular settings because it replaces unavailable smooth linearization by a directly observable infinitesimal dynamics of arcs, secants, or dilations.

Across these contexts, the unifying content of the kinematic tangent cone is precise: tangency is certified by realizable infinitesimal behavior. In singular inclusions this means feasible arcs $x(t)=x_0+th+o(t)$06 (Prusinska et al., 2018); in Euclidean set theory it means secants between pairs of nearby points (Bigolin et al., 2012); in blow-up theories it means asymptotic limits under explicit dilations (Bellettini, 2011). The term therefore designates not merely a cone at a point, but a method for extracting infinitesimal structure from motion.