<?xml version="1.0" encoding="UTF-8"?>

<record version="1" id="1324">
 <title>Minimum-Effort Interception of a Moving Target</title>
 <name>MinimumEffortInterceptionOfAMovingTarget</name>
 <created>2026-09-27 23:10:41</created>
 <modified>2026-09-27 23:10:41</modified>
 <type>Example</type>
<parent id="1200">Calculus of Variations: Endpoints, Natural Boundary Conditions, and Transversality</parent>
 <creator id="1" name="bloftin"/>
 <modifier id="1" name="bloftin"/>
 <comment>adding a julia visual computational resource</comment>
 <author id="1" name="bloftin"/>
 <classification>
	<category scheme="pacs" code="02.30.Xx"/>
	<category scheme="pacs" code="02.30.Sa"/>
	<category scheme="pacs" code="45.20.Jj"/>
 </classification>
 <synonyms>
	<synonym concept="Minimum-Effort Interception of a Moving Target" alias="CV05E3"/>
 </synonyms>
 <keywords>
	<term>calculus of variations</term>
	<term>transversality</term>
	<term>moving endpoint</term>
	<term>terminal constraint</term>
	<term>moving target</term>
	<term>interception</term>
	<term>minimum control effort</term>
	<term>optimal control</term>
	<term>terminal manifold</term>
	<term>costate</term>
	<term>Hamiltonian</term>
 </keywords>
 <preamble>% this is the default PhysicsLibrary preamble.  as your knowledge
% of TeX increases, you will probably want to edit this, but
% it should be fine as is for beginners.

% almost certainly you want these
\usepackage{amssymb}
\usepackage{amsmath}
\usepackage{amsfonts}

% used for TeXing text within eps files
%\usepackage{psfrag}
% need this for including graphics (\includegraphics).
% If you use \includegraphics{figure} with no file extension, PhysicsLibrary
% will prefer figure.png before falling back to figure.eps.
\usepackage{graphicx}
\DeclareGraphicsExtensions{.png,.jpg,.jpeg,.pdf,.eps}
% for neatly defining theorems and propositions
%\usepackage{amsthm}
% making logically defined graphics
%\usepackage{xypic}

% there are many more packages, add them here as you need them

% define commands here</preamble>
 <content>\section*{Minimum-Effort Interception of a Moving Target: Transversality as a Bridge to Optimal Control}

\subsection*{Purpose of this example}

CV05 introduced transversality as the endpoint condition that appears when a
terminal point is not fixed, but is constrained to lie on a curve.  The phrase
``endpoint constrained to a curve'' can sound geometrically abstract.  This
example gives it an immediate physical interpretation: a vehicle must intercept
a target whose position changes with time.

The target trajectory itself is the terminal constraint curve.  The interception
point is not known beforehand.  The optimizer is allowed to choose \emph{when}
and \emph{where} on that curve the interception occurs.  The Euler--Lagrange
equation determines the stationary motion in the interior; the transversality
condition determines how that motion must meet the moving terminal constraint.

The example is deliberately based on the first-order kinematic model

\begin{equation}
x'=u,
\label{eq:kinematics}
\end{equation}

so that the transversality idea is not hidden by higher-order dynamics.  The
control $u$ is a directly commanded velocity.  Later optimal-control models may
replace this by acceleration, thrust, steering angle, or another physical
input.

\section{Problem statement: intercept a moving target}

A one-dimensional vehicle starts at

\begin{equation}
x(0)=0.
\label{eq:initial-condition}
\end{equation}

A target begins a distance $d&gt;0$ ahead and moves away at constant speed
$V&gt;0$:

\begin{equation}
x_{\mathrm{target}}(t)=d+Vt.
\label{eq:target-line}
\end{equation}

The interception time $T$ is not prescribed.  At interception, the vehicle must
satisfy

\begin{equation}
x(T)=d+VT.
\label{eq:terminal-constraint}
\end{equation}

Thus the terminal point $(T,x(T))$ may slide anywhere along the target line in
the $(t,x)$ plane.

We minimize the quadratic control-effort functional

\begin{equation}
J[u]
=\frac12\int_0^T u^2(t)\,dt.
\label{eq:cost-control}
\end{equation}

Using $u=x'$, this becomes the variational problem

\begin{equation}
\boxed{
J[x]
=\frac12\int_0^T x'^{\,2}(t)\,dt
}
\label{eq:cost-functional}
\end{equation}

subject to Eqs.~\eqref{eq:initial-condition} and
\eqref{eq:terminal-constraint}, with $T$ free.

The phrase \emph{minimum effort} here means minimum quadratic control cost.  It
is not, in general, literal fuel consumption or mechanical work.  The example
is intended to expose the variational endpoint structure as cleanly as
possible.\cite{Liberzon,Kirk}

\begin{center}
\includegraphics[width=0.90\textwidth,keepaspectratio]{CV05E3_moving_target_geometry.png}

{\small Figure 1. The terminal point is not a prescribed point.  It may slide
along the moving target trajectory $x=d+Vt$.  For each trial interception time,
the stationary interior path is a straight line.  Transversality selects the
particular point on the target curve at which the optimum occurs.}
\end{center}

\section{Why this is a transversality problem}

In the generic notation of CV05, the independent variable was called $x$ and
the dependent function was $y(x)$.  Here it is more natural to use time $t$ as
the independent variable and position $x(t)$ as the dependent function.  The
translation is therefore

\begin{equation}
x_{\mathrm{CV05}}\longrightarrow t,
\qquad
y_{\mathrm{CV05}}\longrightarrow x(t).
\end{equation}

The terminal point is

\begin{equation}
B=(T,x(T)).
\end{equation}

Because $T$ is free, the endpoint may move horizontally in the $(t,x)$ plane.
Because the vehicle must remain on the target trajectory at interception, the
vertical terminal displacement is not independent of that horizontal motion.

For the target curve

\begin{equation}
x=\psi(t)=d+Vt,
\end{equation}

an allowed endpoint displacement obeys the linearized constraint

\begin{equation}
\boxed{
\delta x_T
=\psi'(T)\,\delta T
=V\,\delta T.
}
\label{eq:tangent-constraint}
\end{equation}

Equation~\eqref{eq:tangent-constraint} says that the endpoint may move, but only
tangent to the target curve.  This is exactly the endpoint geometry for which
CV05 derived a transversality condition.

\section{Deriving the moving-endpoint variation for this problem}

Because the movement of the endpoint is central to this example, we derive the
terminal variation rather than simply quoting the final formula.

Introduce a one-parameter family of nearby curves

\begin{equation}
x_\epsilon(t)=x(t)+\epsilon\eta(t),
\label{eq:family}
\end{equation}

and allow the terminal time to vary as

\begin{equation}
T_\epsilon=T+\epsilon\,\delta T.
\label{eq:moving-time}
\end{equation}

The perturbed functional is

\begin{equation}
J(\epsilon)
=
\int_0^{T_\epsilon}
F\!\left(t,x_\epsilon,x'_\epsilon\right)\,dt,
\label{eq:perturbed-functional}
\end{equation}

where, for the present problem,

\begin{equation}
F(t,x,x')=\frac12x'^{\,2}.
\label{eq:F-definition}
\end{equation}

The first variation is

\begin{equation}
\delta J
=
\left.\frac{dJ}{d\epsilon}\right|_{\epsilon=0}.
\end{equation}

The upper integration limit depends on $\epsilon$, so the Leibniz rule gives

\begin{equation}
\delta J
=
\int_0^T
\left(
F_x\eta+F_{x'}\eta'
\right)dt
+
F(T)\,\delta T.
\label{eq:leibniz-first-var}
\end{equation}

The final term has a simple geometric interpretation.  Moving the upper limit
from $T$ to $T+\delta T$ adds, to first order, a thin strip with width
$\delta T$ and height $F(T)$, hence the contribution $F(T)\delta T$.

Integrate the second term in Eq.~\eqref{eq:leibniz-first-var} by parts:

\begin{align}
\int_0^T F_{x'}\eta'\,dt
&amp;=
\left[F_{x'}\eta\right]_0^T
-
\int_0^T
\frac{d}{dt}F_{x'}\,\eta\,dt.
\end{align}

Therefore

\begin{equation}
\delta J
=
\int_0^T
\left(
F_x-\frac{d}{dt}F_{x'}
\right)\eta\,dt
+
\left[F_{x'}\eta\right]_0^T
+
F(T)\delta T.
\label{eq:first-var-after-ibp}
\end{equation}

The initial point is fixed, so

\begin{equation}
\eta(0)=0.
\end{equation}

For a stationary curve, the interior Euler--Lagrange equation

\begin{equation}
F_x-\frac{d}{dt}F_{x'}=0
\label{eq:EL-general}
\end{equation}

eliminates the integral term.  The terminal contribution becomes

\begin{equation}
\delta J_T
=
F_{x'}(T)\eta(T)
+
F(T)\delta T.
\label{eq:terminal-eta-form}
\end{equation}

\section{The distinction between $\eta(T)$ and the actual endpoint motion}

Equation~\eqref{eq:terminal-eta-form} contains $\eta(T)$, but when the terminal
time moves, $\eta(T)$ is not the same quantity as the actual vertical endpoint
displacement $\delta x_T$.

The perturbed endpoint position is

\begin{equation}
x_\epsilon(T_\epsilon)
=
x_\epsilon(T+\epsilon\delta T).
\end{equation}

Substituting Eq.~\eqref{eq:family},

\begin{equation}
x_\epsilon(T_\epsilon)
=
x(T+\epsilon\delta T)
+
\epsilon\eta(T+\epsilon\delta T).
\end{equation}

Taylor-expand to first order in $\epsilon$:

\begin{align}
x(T+\epsilon\delta T)
&amp;=
x(T)+\epsilon x'(T)\delta T+O(\epsilon^2),
\\
\epsilon\eta(T+\epsilon\delta T)
&amp;=
\epsilon\eta(T)+O(\epsilon^2).
\end{align}

Hence

\begin{equation}
x_\epsilon(T_\epsilon)
=
x(T)
+
\epsilon
\left[
\eta(T)+x'(T)\delta T
\right]
+
O(\epsilon^2).
\end{equation}

The actual first-order terminal displacement is therefore

\begin{equation}
\boxed{
\delta x_T
=
\eta(T)+x'(T)\delta T.
}
\label{eq:endpoint-linearization}
\end{equation}

Equivalently,

\begin{equation}
\boxed{
\eta(T)
=
\delta x_T-x'(T)\delta T.
}
\label{eq:eta-terminal}
\end{equation}

This is the same moving-endpoint linearization developed in CV05.  One term
comes from deforming the curve at fixed $t=T$; the other comes from sliding the
endpoint along the original curve.

Substitute Eq.~\eqref{eq:eta-terminal} into
Eq.~\eqref{eq:terminal-eta-form}:

\begin{align}
\delta J_T
&amp;=
F_{x'}
\left(
\delta x_T-x'\delta T
\right)
+
F\delta T
\\
&amp;=
F_{x'}\delta x_T
+
\left(
F-x' F_{x'}
\right)\delta T.
\end{align}

Thus the universal terminal boundary form in the $(t,x)$ plane is

\begin{equation}
\boxed{
\delta J_T
=
F_{x'}\,\delta x_T
+
\left(F-x' F_{x'}\right)\delta T.
}
\label{eq:universal-boundary}
\end{equation}

\section{Apply the target-curve constraint}

The endpoint must remain on the moving target line.  Therefore
Eq.~\eqref{eq:tangent-constraint} gives

\begin{equation}
\delta x_T=V\delta T.
\end{equation}

Substitute into Eq.~\eqref{eq:universal-boundary}:

\begin{align}
\delta J_T
&amp;=
F_{x'}V\delta T
+
\left(F-x'F_{x'}\right)\delta T
\\
&amp;=
\left[
F+(V-x')F_{x'}
\right]\delta T.
\end{align}

The interception point is free to move along the target trajectory, so
$\delta T$ is arbitrary.  Stationarity therefore requires

\begin{equation}
\boxed{
F+(V-x')F_{x'}=0
\qquad\text{at }t=T.
}
\label{eq:transversality-general}
\end{equation}

This is the transversality condition for this moving-target problem.

The key contrast with a natural boundary condition is worth emphasizing.  A
natural boundary condition appears when a terminal coordinate is freely
variable by itself.  Here the terminal variations are \emph{linked}:

\begin{equation}
\delta x_T=V\delta T.
\end{equation}

The endpoint has one degree of freedom, but that freedom lies tangent to the
target curve.  Transversality is the stationarity condition associated with
that constrained endpoint motion.

\section{Euler--Lagrange equation in the interior}

For

\begin{equation}
F=\frac12x'^{\,2},
\end{equation}

we have

\begin{equation}
F_x=0,
\qquad
F_{x'}=x'.
\end{equation}

The Euler--Lagrange equation is therefore

\begin{equation}
0-\frac{d}{dt}(x')=0,
\end{equation}

or

\begin{equation}
\boxed{
x''=0.
}
\label{eq:ddot-zero}
\end{equation}

Integrating twice,

\begin{equation}
x(t)=C_0+C_1t.
\end{equation}

The fixed initial condition $x(0)=0$ gives $C_0=0$, so

\begin{equation}
\boxed{
x(t)=C_1t,
\qquad
u(t)=x'(t)=C_1.
}
\label{eq:constant-control}
\end{equation}

Thus the stationary interior trajectory uses a constant velocity command.  The
remaining question is: \emph{which constant velocity, and which interception
point on the target curve?}  That is precisely what the terminal constraint and
transversality condition determine.

\section{Use transversality to determine the optimal control}

Substitute

\begin{equation}
F=\frac12x'^{\,2},
\qquad
F_{x'}=x'
\end{equation}

into Eq.~\eqref{eq:transversality-general}:

\begin{equation}
\frac12x'^{\,2}
+
(V-x')x'
=0.
\end{equation}

Expand:

\begin{equation}
\frac12x'^{\,2}
+Vx'
-x'^{\,2}
=0,
\end{equation}

so

\begin{equation}
Vx'
-\frac12x'^{\,2}
=0.
\end{equation}

Factor:

\begin{equation}
x'
\left(
V-\frac12x'
\right)
=0.
\end{equation}

The root $x'=0$ cannot intercept a target that begins at $d&gt;0$.  The
nontrivial stationary solution is therefore

\begin{equation}
\boxed{
x'=u^*=2V.
}
\label{eq:u-optimal}
\end{equation}

The terminal constraint now determines $T$:

\begin{equation}
2VT=d+VT.
\end{equation}

Hence

\begin{equation}
\boxed{
T^*=\frac{d}{V}.
}
\label{eq:T-optimal}
\end{equation}

The interception position is

\begin{equation}
\boxed{
x(T^*)=2d.
}
\label{eq:x-intercept}
\end{equation}

The target travels a distance $d$ during the optimal interception interval, so
it moves from $x=d$ to $x=2d$.  The vehicle travels twice that distance from the
origin, at twice the target speed.

\section{What does the solution mean physically?}

The result

\begin{equation}
u^*=2V,
\qquad
T^*=\frac{d}{V}
\end{equation}

has a useful physical interpretation.

If the vehicle tries to intercept too early, the target has not moved very far,
but the vehicle must cover the initial gap $d$ in a very short time.  The
required command $u$ is large, and the quadratic penalty $u^2$ makes that
expensive.

If the vehicle waits too long, the required excess speed above the target speed
is smaller, but the target continues to move away and the control must be
applied over a longer interval.  The accumulated cost again becomes large.

The optimum balances these two effects.  At $T=d/V$, the target has moved
exactly one initial-gap distance $d$.  The vehicle has traveled $2d$, and the
relative closing speed is

\begin{equation}
u^*-V=V.
\end{equation}

Thus the time needed to close the original gap is

\begin{equation}
\frac{d}{u^*-V}
=
\frac{d}{V}
=T^*.
\end{equation}

The transversality condition has therefore selected both a terminal point and a
control level with a simple physical meaning.

\begin{center}
\includegraphics[width=0.82\textwidth,keepaspectratio]{CV05E3_optimal_control_profile.png}

{\small Figure 2. The stationary interior solution has constant control because
$x''=0$.  Transversality selects the particular level $u^*=2V$.  The target
moves at speed $V$, so the optimal vehicle command is twice the target speed.}
\end{center}

\section{An independent check: optimize directly over interception time}

There is a useful way to verify the transversality result without using the
moving-endpoint formula.

Suppose temporarily that a trial interception time $T$ has been chosen.  The
terminal point is then fixed at

\begin{equation}
x(T)=d+VT.
\end{equation}

For this fixed $T$, the Euler--Lagrange equation gives a constant-speed path
from $0$ to $d+VT$:

\begin{equation}
u_T
=\frac{d+VT}{T}
=V+\frac{d}{T}.
\label{eq:u-of-T}
\end{equation}

The corresponding cost is

\begin{align}
J(T)
&amp;=\frac12u_T^2T
\\
&amp;=\frac12
\left(V+\frac{d}{T}\right)^2T
\\
&amp;=\boxed{
\frac{d^2}{2T}+dV+\frac{V^2T}{2}
}.
\label{eq:J-of-T}
\end{align}

Differentiate:

\begin{equation}
\frac{dJ}{dT}
=-\frac{d^2}{2T^2}+\frac{V^2}{2}.
\end{equation}

Stationarity gives

\begin{equation}
\frac{d^2}{T^2}=V^2.
\end{equation}

Because $T&gt;0$, $d&gt;0$, and $V&gt;0$,

\begin{equation}
T=\frac{d}{V},
\end{equation}

exactly as obtained from transversality.

Moreover,

\begin{equation}
\frac{d^2J}{dT^2}
=\frac{d^2}{T^3}&gt;0,
\end{equation}

so this stationary value is a strict minimum with respect to the interception
time.

At the optimum, the two $T$-dependent contributions in Eq.~\eqref{eq:J-of-T}
are equal:

\begin{equation}
\frac{d^2}{2T^*}
=
\frac{V^2T^*}{2}
=\frac{dV}{2}.
\end{equation}

This equality gives another interpretation of the optimum: the cost associated
with closing the initial spatial gap and the cost associated with the target's
continued motion are balanced.

\begin{center}
\includegraphics[width=0.82\textwidth,keepaspectratio]{CV05E3_cost_vs_intercept_time.png}

{\small Figure 3. Normalized cost versus normalized interception time.  Very
early interception is expensive because the required speed is large; very late
interception is expensive because the control is maintained while the target
continues to move away.  The minimum occurs at $VT/d=1$.}
\end{center}

\section{Minimum cost}

Substitute Eqs.~\eqref{eq:u-optimal} and \eqref{eq:T-optimal} into the objective:

\begin{align}
J^*
&amp;=\frac12\int_0^{d/V}(2V)^2\,dt
\\
&amp;=\frac12(4V^2)\frac{d}{V}.
\end{align}

Therefore

\begin{equation}
\boxed{
J^*=2dV.
}
\label{eq:J-optimal}
\end{equation}

The units are those of the mathematical quadratic control cost.  Since $u$ is a
velocity command here,

\begin{equation}
[J]=\frac{\mathrm{length}^2}{\mathrm{time}}.
\end{equation}

This should not be interpreted automatically as mechanical energy.

\section{Numerical example}

Take

\begin{equation}
d=100\ \mathrm{m},
\qquad
V=10\ \mathrm{m/s}.
\end{equation}

The transversality solution gives

\begin{equation}
u^*=2V=20\ \mathrm{m/s},
\end{equation}

and

\begin{equation}
T^*=\frac{100}{10}=10\ \mathrm{s}.
\end{equation}

The target begins at $100\ \mathrm{m}$ and moves another $100\ \mathrm{m}$ during
those ten seconds, so the interception occurs at

\begin{equation}
x(T^*)=200\ \mathrm{m}.
\end{equation}

The minimum quadratic control cost is

\begin{equation}
J^*=2(100)(10)=2000\ \mathrm{m^2/s}.
\end{equation}

The vehicle's relative closing speed is

\begin{equation}
20-10=10\ \mathrm{m/s},
\end{equation}

so closing the initial $100\ \mathrm{m}$ gap takes exactly $10\ \mathrm{s}$.

\section{Geometric meaning of transversality}

Define the momentum-like endpoint coefficient

\begin{equation}
p=F_{x'}
\end{equation}

and the Hamiltonian-like quantity

\begin{equation}
\mathcal H
=x'\,p-F.
\end{equation}

Then the terminal boundary form, Eq.~\eqref{eq:universal-boundary}, becomes

\begin{equation}
\boxed{
\delta J_T
=p\,\delta x_T
-\mathcal H\,\delta T.
}
\label{eq:pH-boundary}
\end{equation}

The terminal target curve has tangent displacement

\begin{equation}
(\delta T,\delta x_T)
=\delta T(1,V).
\end{equation}

Equation~\eqref{eq:pH-boundary} can therefore be written as the dot product

\begin{equation}
\delta J_T
=
(-\mathcal H,p)
\cdot
(\delta T,\delta x_T).
\end{equation}

Stationarity requires this dot product to vanish for every allowed tangent
motion.  Hence

\begin{equation}
\boxed{
(-\mathcal H,p)
\perp
(1,V).
}
\label{eq:normal-condition}
\end{equation}

The boundary covector is normal to the terminal constraint curve.  This is the
geometric content of the word \emph{transversality}.

For the present integrand,

\begin{equation}
p=x'=u,
\qquad
\mathcal H
=u^2-\frac12u^2
=\frac12u^2.
\end{equation}

Thus Eq.~\eqref{eq:normal-condition} gives

\begin{equation}
-\frac12u^2+uV=0,
\end{equation}

which again yields the nontrivial solution $u=2V$.

\begin{center}
\includegraphics[width=0.76\textwidth,keepaspectratio]{CV05E3_transversality_geometry.png}

{\small Figure 4. At the selected terminal point, allowed endpoint motion is
tangent to the target curve.  The variational boundary covector
$(-\mathcal H,p)$ must annihilate that tangent motion and is therefore normal to
the target curve.  This orthogonality in endpoint-variation space is the
geometric meaning of transversality.}
\end{center}

\section{The same result in optimal-control language}

Now rewrite the problem exactly as an elementary optimal-control problem:

\begin{equation}
x'=u,
\qquad
J=\int_0^T \frac12u^2\,dt.
\end{equation}

Use the minimization-convention control Hamiltonian

\begin{equation}
H_c
=\frac12u^2+\lambda u.
\label{eq:control-H}
\end{equation}

Stationarity with respect to $u$ gives

\begin{equation}
\frac{\partial H_c}{\partial u}
=u+\lambda=0,
\end{equation}

so

\begin{equation}
\boxed{
u=-\lambda.
}
\label{eq:u-lambda}
\end{equation}

The costate equation is

\begin{equation}
\lambda'
=-\frac{\partial H_c}{\partial x}=0,
\end{equation}

so the costate is constant, and therefore the optimal control is constant.  This
is the same conclusion obtained from the Euler--Lagrange equation
$x''=0$.

The terminal condition is not a fixed terminal state.  It is the terminal
manifold

\begin{equation}
g(T,x(T))
=x(T)-d-VT
=0.
\label{eq:terminal-manifold}
\end{equation}

Introduce a terminal multiplier $\nu$.  The standard transversality conditions
for this equality constraint, with no terminal cost, are

\begin{equation}
\lambda(T)=\nu g_x,
\label{eq:lambda-terminal}
\end{equation}

and

\begin{equation}
H_c(T)+\nu g_T=0.
\label{eq:H-terminal}
\end{equation}

For Eq.~\eqref{eq:terminal-manifold},

\begin{equation}
g_x=1,
\qquad
g_T=-V.
\end{equation}

Therefore

\begin{equation}
\lambda(T)=\nu,
\end{equation}

and

\begin{equation}
H_c(T)-\nu V=0.
\end{equation}

From Eq.~\eqref{eq:u-lambda}, $\nu=\lambda=-u$.  On the stationary control,

\begin{equation}
H_c
=\frac12u^2-u^2
=-\frac12u^2.
\end{equation}

Thus the terminal condition becomes

\begin{equation}
-\frac12u^2+uV=0,
\end{equation}

and again

\begin{equation}
u^*=2V.
\end{equation}

The same terminal geometry has therefore appeared in three equivalent forms:

\begin{equation}
\boxed{
\begin{array}{c}
\text{CV endpoint condition}
\\[3pt]
F+(V-x')F_{x'}=0

\end{array}
}
\end{equation}

\begin{equation}
\boxed{
(-\mathcal H,p)\perp(1,V)
}
\end{equation}

and

\begin{equation}
\boxed{
\lambda(T)=\nu g_x,
\qquad
H_c(T)+\nu g_T=0.
}
\end{equation}

The first is classical calculus of variations, the second is its geometric
boundary interpretation, and the third is the terminal transversality language
of optimal control.\cite{Liberzon,Kirk}

\section{Natural boundary condition versus transversality}

CV05E2 considered a free terminal state coordinate.  This article considers a
terminal point constrained to a curve.  The distinction can be summarized as
follows.

\begin{center}
\begin{tabular}{|p{0.26\textwidth}|p{0.30\textwidth}|p{0.32\textwidth}|}
\hline
\textbf{Endpoint type} &amp; \textbf{Allowed variation} &amp; \textbf{Boundary consequence} \\
\hline
Fixed terminal coordinate &amp; Variation of that coordinate is zero &amp; No new boundary equation from that coordinate \\
\hline
Free terminal coordinate &amp; Variation is arbitrary and independent &amp; Its conjugate boundary coefficient must vanish \\
\hline
Endpoint constrained to a curve &amp; Endpoint variations are nonzero but linked by the curve tangent &amp; A linear combination of boundary coefficients must vanish: transversality \\
\hline
\end{tabular}
\end{center}

In the present example,

\begin{equation}
\delta x_T=V\delta T
\end{equation}

is the crucial relation.  Neither $\delta x_T$ nor $\delta T$ is forced to zero,
but they cannot vary independently.  That is why the terminal condition is not
simply $F_{x'}=0$ or $F-x'F_{x'}=0$.  Instead the two coefficients
combine according to the tangent direction of the target curve.

\section{Generalization to a nonlinear moving target}

Suppose the target trajectory is a smooth nonlinear function

\begin{equation}
x_{\mathrm{target}}(t)=\psi(t).
\end{equation}

Then an allowed terminal displacement satisfies

\begin{equation}
\delta x_T=\psi'(T)\delta T.
\end{equation}

The transversality condition becomes

\begin{equation}
\boxed{
F+
\left(\psi'(T)-x'(T)\right)F_{x'}
=0.
}
\end{equation}

For $F=\tfrac12x'^2$, the nontrivial terminal condition reduces to

\begin{equation}
\boxed{
x'(T)=2\psi'(T).
}
\end{equation}

The interior Euler--Lagrange equation still requires $x'$ to be constant in
this simple single-integrator model, so the interception time must satisfy both

\begin{equation}
x(T)=\psi(T)
\end{equation}

and

\begin{equation}
x'=2\psi'(T).
\end{equation}

For a nonlinear target path, these equations may need to be solved numerically.
This is already very close in spirit to practical trajectory optimization.

\section{What should be carried forward into optimal control?}

This example provides several ideas that recur in more advanced trajectory
optimization:

\begin{enumerate}
\item The terminal state need not be a prescribed point.  It may lie on a
      terminal curve, surface, or higher-dimensional manifold.
\item The Euler--Lagrange equations govern the stationary trajectory in the
      interior, but they do not by themselves determine which terminal point on
      the constraint should be selected.
\item Allowed endpoint variations are tangent to the terminal constraint.
\item The variational boundary covector must annihilate every allowed tangent
      displacement; geometrically, it is normal to the terminal constraint.
\item In optimal control, the same normality condition is expressed through
      terminal costates and multipliers.
\item A transversality condition is a necessary stationarity condition.  It
      does not by itself prove that the stationary trajectory is a minimum.
\end{enumerate}

For the moving-target problem, the conceptual chain is

\begin{equation}
\boxed{
\begin{array}{c}
\text{moving target}
\;\longrightarrow\;
\text{terminal curve}
\;\longrightarrow\;
\text{linked endpoint variations}
\\[4pt]
\phantom{\text{moving target}}
\;\longrightarrow\;
\text{transversality}
\;\longrightarrow\;
\text{terminal manifold in optimal control}
\end{array}
}
\end{equation}

\section*{Summary}

A kinematic vehicle obeying

\begin{equation*}
x'=u
\end{equation*}

must intercept a target

\begin{equation*}
x_{\mathrm{target}}(t)=d+Vt
\end{equation*}

while minimizing

\begin{equation*}
J=\frac12\int_0^T u^2\,dt.
\end{equation*}

The final time is free, and the final point must lie on the target trajectory.
The moving-endpoint first variation produces

\begin{equation*}
\delta J_T
=F_{x'}\delta x_T
+
\left(F-x'F_{x'}\right)\delta T.
\end{equation*}

Because the target curve imposes

\begin{equation*}
\delta x_T=V\delta T,
\end{equation*}

stationarity requires

\begin{equation*}
F+(V-x')F_{x'}=0.
\end{equation*}

For

\begin{equation*}
F=\frac12x'^2,
\end{equation*}

Euler--Lagrange gives constant $x'$, while transversality selects

\begin{equation*}
\boxed{
u^*=2V.
}
\end{equation*}

The terminal constraint then gives

\begin{equation*}
\boxed{
T^*=\frac{d}{V},
\qquad
x(T^*)=2d,
\qquad
J^*=2dV.
}
\end{equation*}

The example shows concretely what it means for an endpoint to be ``constrained
to a curve.''  The terminal point may move, but only tangent to the target
trajectory.  Transversality supplies the boundary equation associated with
that constrained motion, and the same structure becomes a terminal-manifold
condition in optimal control.

\section*{References}

\begin{thebibliography}{9}

\bibitem{GelfandFomin}
I. M. Gelfand and S. V. Fomin,
\emph{Calculus of Variations},
Prentice-Hall, 1963; reprinted by Dover Publications.

\bibitem{VanBrunt}
B. van Brunt,
\emph{The Calculus of Variations},
Springer, 2004.

\bibitem{Liberzon}
D. Liberzon,
\emph{Calculus of Variations and Optimal Control Theory: A Concise Introduction},
Princeton University Press, 2012.

\bibitem{Kirk}
D. E. Kirk,
\emph{Optimal Control Theory: An Introduction},
Prentice-Hall, 1970.

\end{thebibliography}</content>
</record>
