Namespaces
Variants
Actions

Difference between revisions of "Transversality condition"

From Encyclopedia of Mathematics
Jump to: navigation, search
(Importing text file)
 
m
 
(2 intermediate revisions by 2 users not shown)
Line 3: Line 3:
 
For the simplest problem in variational calculus with variable end-points,
 
For the simplest problem in variational calculus with variable end-points,
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t0940101.png" /></td> </tr></table>
+
$$
 +
J(x) = \int_{t_1}^{t_2} F(t, x, \dot x) \, dt,
 +
$$
  
 
in which the point
 
in which the point
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t0940102.png" /></td> </tr></table>
+
$$
 +
(t_1, x(t_1), t_2, x(t_2)) = (t_1, x_1, t_2, x_2)
 +
$$
  
 
is not fixed but can belong to a certain manifold, the transversality condition can be written in the form of the relation
 
is not fixed but can belong to a certain manifold, the transversality condition can be written in the form of the relation
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t0940103.png" /></td> <td valign="top" style="width:5%;text-align:right;">(1)</td></tr></table>
+
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;">
 +
$$
 +
[(F - \dot x F_{\dot x}) \ dt + F_{\dot x} \ dx]_1^2 = 0,
 +
$$
 +
</td> <td valign="top" style="width:5%;text-align:right;">(1)</td></tr></table>
  
which must be satisfied for any values of the tangent differentials <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t0940104.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t0940105.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t0940106.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t0940107.png" /> of the boundary condition manifold.
+
which must be satisfied for any values of the tangent differentials $d t_1$, $d x_1$, $dt_2$, $d x_2$ of the boundary condition manifold.
  
If the left- and right-hand end-points of the extremal can be displaced along prescribed curves <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t0940108.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t0940109.png" />, then since
+
If the left- and right-hand end-points of the extremal can be displaced along prescribed curves $x = \phi_1(t)$ and $x = \phi_2(t)$, then since
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401010.png" /></td> </tr></table>
+
$$d x_1 = \dot \phi_1(t) \, dt_1, \qquad d x_2 = \dot \phi_2(t) \, dt_2$$
  
and the variations of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401011.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401012.png" /> are independent, (1) implies
+
and the variations of $dt_1$ and $dt_2$ are independent, (1) implies
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401013.png" /></td> <td valign="top" style="width:5%;text-align:right;">(2)</td></tr></table>
+
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;">
 +
$$
 +
\left.
 +
\begin{array}{l}
 +
F(t_1, x_1, \dot x_1) - [\dot \phi_1(t_1) - \dot x_1] F_{\dot x} (t_1, x_1, \dot x_1) = 0, \\
 +
F(t_2, x_2, \dot x_2) - [\dot \phi_2(t_2) - \dot x_2] F_{\dot x} (t_2, x_2, \dot x_2) = 0.
 +
\end{array}
 +
\right\}
 +
$$
 +
</td> <td valign="top" style="width:5%;text-align:right;">(2)</td></tr></table>
  
If the equations of the curves along which the left- and right-hand end-points are displaced are given in implicit form, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401014.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401015.png" />, then the transversality condition (1) can be written in the form
+
If the equations of the curves along which the left- and right-hand end-points are displaced are given in implicit form, $\omega_1(t,x) = 0$ and $\omega_2(t,x) = 0$, then the transversality condition (1) can be written in the form
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401016.png" /></td> <td valign="top" style="width:5%;text-align:right;">(3)</td></tr></table>
+
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;">
 +
$$
 +
\left.
 +
\begin{array}{ll}
 +
\dfrac{F - \dot x F_{\dot x}}{w_{1t}} = \dfrac{F{\dot x}}{\omega_{1x}} & \text{at the left-hand end-point,} \\
 +
\dfrac{F - \dot x F_{\dot x}}{w_{2t}} = \dfrac{F{\dot x}}{\omega_{2x}} & \text{at the right-hand end-point}
 +
\end{array}
 +
\right\}
 +
$$
 +
</td> <td valign="top" style="width:5%;text-align:right;">(3)</td></tr></table>
  
If there are no constraints on one of the end-points, then at this end-point, by virtue of the independence of the respective tangent differentials <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401017.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401018.png" />, the transversality condition takes the form
+
If there are no constraints on one of the end-points, then at this end-point, by virtue of the independence of the respective tangent differentials $dt$ and $dx$, the transversality condition takes the form
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401019.png" /></td> <td valign="top" style="width:5%;text-align:right;">(4)</td></tr></table>
+
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;">
 +
$$
 +
F = 0, \qquad F_{\dot x} = 0.
 +
$$
 +
</td> <td valign="top" style="width:5%;text-align:right;">(4)</td></tr></table>
  
 
The relations (2), (3), (4) are called transversality conditions.
 
The relations (2), (3), (4) are called transversality conditions.
Line 35: Line 65:
 
Below, the transversality condition is given in the more general case of the variational problem for a conditional extremum. Consider the [[Bolza problem|Bolza problem]], i.e. the problem of minimizing a functional
 
Below, the transversality condition is given in the more general case of the variational problem for a conditional extremum. Consider the [[Bolza problem|Bolza problem]], i.e. the problem of minimizing a functional
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401020.png" /></td> <td valign="top" style="width:5%;text-align:right;">(5)</td></tr></table>
+
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;">
 +
$$
 +
\begin{gathered}
 +
J(x) = \int_{t_1}^{t_2} f(t,x,\dot x) dt + g(t_1, x(t_1), t_2, x(t_2)),  \\
 +
f:\R \times \R^n \times \R^n \to \R, \qquad g:\R \times \R^n \times \R \times \R^n \to \R,
 +
\end{gathered}
 +
$$
 +
</td> <td valign="top" style="width:5%;text-align:right;">(5)</td></tr></table>
  
 
in the presence of differential constraints of equality type,
 
in the presence of differential constraints of equality type,
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401021.png" /></td> <td valign="top" style="width:5%;text-align:right;">(6)</td></tr></table>
+
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;">
 +
$$
 +
\phi(t,x,\dot x) = 0, \qquad \phi: \R\times \R^n\times \R^n\to\R^m, \quad m<n,
 +
$$
 +
</td> <td valign="top" style="width:5%;text-align:right;">(6)</td></tr></table>
  
 
and boundary conditions
 
and boundary conditions
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401022.png" /></td> <td valign="top" style="width:5%;text-align:right;">(7)</td></tr></table>
+
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;">
 +
$$
 +
\begin{gathered}
 +
\psi(t_1, x(t_1), t_2, x(t_2)) = 0,\\
 +
\psi: \R \times \R^n\times\R\times\R^n\to\R^p, \qquad p\le 2n+2.
 +
\end{gathered}
 +
$$
 +
</td> <td valign="top" style="width:5%;text-align:right;">(7)</td></tr></table>
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401023.png" /></td> </tr></table>
+
When $p < 2n+2$ in this problem, the end-points $(t_1, x_1^1, \ldots, x_1^n)$ and $(t_2, x_2^1, \ldots, x_2^n)$ of the extremal are not fixed, but can be displaced along given hypersurfaces $\psi_\mu = 0$, $\mu = 1, \ldots, p$.
  
When <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401024.png" /> in this problem, the end-points <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401025.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401026.png" /> of the extremal are not fixed, but can be displaced along given hypersurfaces <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401027.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401028.png" />.
+
In accordance with the transversality condition, there exist constants (Lagrange multipliers) $e_\mu$, $\mu=1,\ldots,p$, as well as multipliers $\lambda_0$ and $\lambda_i(t)$, $i=1,\ldots,m$, such that, in addition to the boundary conditions (7), the following relation holds at the end-points of the extremal:
  
In accordance with the transversality condition, there exist constants (Lagrange multipliers) <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401029.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401030.png" />, as well as multipliers <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401031.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401032.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401033.png" />, such that, in addition to the boundary conditions (7), the following relation holds at the end-points of the extremal:
+
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;">
 
+
$$
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401034.png" /></td> <td valign="top" style="width:5%;text-align:right;">(8)</td></tr></table>
+
\left[
 
+
\left( F - \sum_{i=1}^n \dot x^i F_{\dot x^i}\right) \ dt +
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401035.png" /></td> </tr></table>
+
\sum_{i=1}^n F_{\dot x^i} \ dx^i
 +
\right]_1^2 + \lambda_0 \ dg + \sum_{\mu=1}^p e_\mu \ d\psi_\mu =0,
 +
$$
 +
</td> <td valign="top" style="width:5%;text-align:right;">(8)</td></tr></table>
  
 
for any choice of tangent differentials
 
for any choice of tangent differentials
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401036.png" /></td> <td valign="top" style="width:5%;text-align:right;">(9)</td></tr></table>
+
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;">
 +
$$
 +
dt_1, dx_1^i, dt_2, dx_2^i, \qquad i = 1,\ldots, n,
 +
$$
 +
</td> <td valign="top" style="width:5%;text-align:right;">(9)</td></tr></table>
  
of the manifold defined by (7). In (8), <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401037.png" /> denotes the expression
+
of the manifold defined by (7). In (8), $F$ denotes the expression
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401038.png" /></td> <td valign="top" style="width:5%;text-align:right;">(10)</td></tr></table>
+
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;">
 +
$$
 +
F = F(t,x,\dot x, \lambda) = \lambda_0 f(t,x,\dot x) + \sum_{i=1}^m \lambda_i(t) \phi_i(t, x, \dot x).
 +
$$
 +
</td> <td valign="top" style="width:5%;text-align:right;">(10)</td></tr></table>
  
In the majority of practical problems, the Lagrange multipliers are normalized by setting <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401039.png" /> (the value <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401040.png" /> corresponds to an abnormal case, see [[#References|[1]]]). The multipliers <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401041.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401042.png" />, are determined together with the <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401043.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401044.png" />, from the solution of the Euler system of differential equations
+
In the majority of practical problems, the Lagrange multipliers are normalized by setting $\lambda_0=1$ (the value $\lambda_0=0$ corresponds to an abnormal case, see [[#References|[1]]]). The multipliers $\lambda_i(t)$, $i=1,\ldots,m$, are determined together with the $x^i(t)$, $i=1,\ldots,m$, from the solution of the Euler system of differential equations
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401045.png" /></td> <td valign="top" style="width:5%;text-align:right;">(11)</td></tr></table>
+
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;">
 +
$$
 +
F_{x^i} - \frac{d}{dt} F_{\dot x^i} = 0, \qquad i = 1,\ldots,n,
 +
$$
 +
</td> <td valign="top" style="width:5%;text-align:right;">(11)</td></tr></table>
  
and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401046.png" /> equations of the form (6):
+
and $m$ equations of the form (6):
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401047.png" /></td> </tr></table>
+
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;">
 +
$$
 +
\phi_i(t,x,\dot x) = 0, \qquad i=1,\dots, m.
 +
$$
 +
</td> </tr></table>
  
The general solution of this system of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401048.png" /> second-order differential equations and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401049.png" /> first-order differential equations in the <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401050.png" /> unknown functions <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401051.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401052.png" />, and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401053.png" />, <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401054.png" />, depends on <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401055.png" /> arbitrary constants. In fact, if one sets
+
The general solution of this system of $n$ second-order differential equations and $m$ first-order differential equations in the $n+m$ unknown functions $x^i(t)$, $i=1,\ldots,n$, and $\lambda_i(t)$, $i=1,\ldots,m$, depends on $2n$ arbitrary constants. In fact, if one sets
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401056.png" /></td> <td valign="top" style="width:5%;text-align:right;">(12)</td></tr></table>
+
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;">
 +
$$
 +
\dot x^i = u_i, \qquad i = 1,\ldots,n, 
 +
$$
 +
</td> <td valign="top" style="width:5%;text-align:right;">(12)</td></tr></table>
  
then one obtains a system (11), (12) of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401057.png" /> first-order differential equations and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401058.png" /> finite relations
+
then one obtains a system (11), (12) of $2n$ first-order differential equations and $m$ finite relations
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401059.png" /></td> <td valign="top" style="width:5%;text-align:right;">(13)</td></tr></table>
+
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;">
 +
$$
 +
\phi_i(t,x,u) = 0, \qquad i = 1,\ldots,m.
 +
$$
 +
</td> <td valign="top" style="width:5%;text-align:right;">(13)</td></tr></table>
  
Using (13), some <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401060.png" /> of the functions <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401061.png" /> can be expressed in terms of the others (under the hypothesis that the corresponding functional determinant does not vanish) and, on substituting these in (11), (12), one obtains a system of <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401062.png" /> first-order differential equations with <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401063.png" /> unknown functions, the general solution of which depends on <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401064.png" /> arbitrary constants. Along with the values <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401065.png" /> and <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401066.png" />, this gives <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401067.png" /> arbitrary constants, determining the solution of the variational problem (5)–(7). One then obtains by means of the transversality condition the correct number of equations enabling one to determine these arbitrary constants.
+
Using (13), some $m$ of the functions $u_i$ can be expressed in terms of the others (under the hypothesis that the corresponding functional determinant does not vanish) and, on substituting these in (11), (12), one obtains a system of $2n$ first-order differential equations with $2n$ unknown functions, the general solution of which depends on $2n$ arbitrary constants. Along with the values $t_1$ and $t_2$, this gives $2n+2$ arbitrary constants, determining the solution of the variational problem (5)–(7). One then obtains by means of the transversality condition the correct number of equations enabling one to determine these arbitrary constants.
  
 
In problems of optimal control and in the [[Pontryagin maximum principle|Pontryagin maximum principle]], the necessary transversality condition is written similarly to (8), only instead of
 
In problems of optimal control and in the [[Pontryagin maximum principle|Pontryagin maximum principle]], the necessary transversality condition is written similarly to (8), only instead of
  
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;"><img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401068.png" /></td> </tr></table>
+
<table class="eq" style="width:100%;"> <tr><td valign="top" style="width:94%;text-align:center;">
 +
$$
 +
F - \sum_{i=1}^n \dot x^i F_{\dot x^i} \qquad \text{and} \qquad F_{\dot x^i}
 +
$$
 +
</td> </tr></table>
  
one has to substitute in (8) the Hamiltonian <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401069.png" />, taken with the opposite sign, and the conjugate variables <img align="absmiddle" border="0" src="https://www.encyclopediaofmath.org/legacyimages/t/t094/t094010/t09401070.png" />.
+
one has to substitute in (8) the Hamiltonian $H$, taken with the opposite sign, and the conjugate variables $\psi_i$.
  
 
The necessary transversality condition gives the missing boundary conditions for obtaining a closed boundary value problem to which the solution of the variational problem with variable end-points reduces.
 
The necessary transversality condition gives the missing boundary conditions for obtaining a closed boundary value problem to which the solution of the variational problem with variable end-points reduces.
  
 
====References====
 
====References====
<table><TR><TD valign="top">[1]</TD> <TD valign="top">  G.A. Bliss,  "Lectures on the calculus of variations" , Chicago Univ. Press  (1947)</TD></TR><TR><TD valign="top">[2]</TD> <TD valign="top">  M.A. Lavrent'ev,  L.A. Lyusternik,  "A course in variational calculus" , Moscow-Leningrad  (1950)  (In Russian)</TD></TR></table>
+
<table>
 
+
  <TR><TD valign="top">[1]</TD> <TD valign="top">  G.A. Bliss,  "Lectures on the calculus of variations" , Chicago Univ. Press  (1947)</TD></TR>
 
+
  <TR><TD valign="top">[2]</TD> <TD valign="top">  M.A. Lavrent'ev,  L.A. Lyusternik,  "A course in variational calculus" , Moscow-Leningrad  (1950)  (In Russian)</TD></TR>
  
====Comments====
+
  <TR><TD valign="top">[a1]</TD> <TD valign="top">  L. Cesari,  "Optimization - Theory and applications" , Springer  (1983)</TD></TR>
 +
  <TR><TD valign="top">[a2]</TD> <TD valign="top">  L.D. Berkovitz,  "Optimal control theory" , Springer  (1974)</TD></TR>
 +
  <TR><TD valign="top">[a3]</TD> <TD valign="top">  L.E. [L.E. El'sgol'ts] Elsgolc,  "Calculus of variations" , Pergamon  (1961)  (Translated from Russian)</TD></TR>
 +
  <TR><TD valign="top">[a4]</TD> <TD valign="top">  R.H. Rishel,  "Deterministic and stochastic optimal control" , Springer  (1975)</TD></TR>
 +
</table>
  
 
+
{{TEX|done}}
====References====
 
<table><TR><TD valign="top">[a1]</TD> <TD valign="top">  L. Cesari,  "Optimization - Theory and applications" , Springer  (1983)</TD></TR><TR><TD valign="top">[a2]</TD> <TD valign="top">  L.D. Berkovitz,  "Optimal control theory" , Springer  (1974)</TD></TR><TR><TD valign="top">[a3]</TD> <TD valign="top">  L.E. [L.E. El'sgol'ts] Elsgolc,  "Calculus of variations" , Pergamon  (1961)  (Translated from Russian)</TD></TR><TR><TD valign="top">[a4]</TD> <TD valign="top">  R.H. Rishel,  "Deterministic and stochastic optimal control" , Springer  (1975)</TD></TR></table>
 

Latest revision as of 08:59, 13 February 2024

A necessary condition for optimality in variational problems with variable end-points. The arbitrary constants on which the solution of the Euler equation depends are determined by means of the tranversality condition. The transversality condition is a necessary condition for the vanishing of the first variation of a functional.

For the simplest problem in variational calculus with variable end-points,

$$ J(x) = \int_{t_1}^{t_2} F(t, x, \dot x) \, dt, $$

in which the point

$$ (t_1, x(t_1), t_2, x(t_2)) = (t_1, x_1, t_2, x_2) $$

is not fixed but can belong to a certain manifold, the transversality condition can be written in the form of the relation

$$ [(F - \dot x F_{\dot x}) \ dt + F_{\dot x} \ dx]_1^2 = 0, $$

(1)

which must be satisfied for any values of the tangent differentials $d t_1$, $d x_1$, $dt_2$, $d x_2$ of the boundary condition manifold.

If the left- and right-hand end-points of the extremal can be displaced along prescribed curves $x = \phi_1(t)$ and $x = \phi_2(t)$, then since

$$d x_1 = \dot \phi_1(t) \, dt_1, \qquad d x_2 = \dot \phi_2(t) \, dt_2$$

and the variations of $dt_1$ and $dt_2$ are independent, (1) implies

$$ \left. \begin{array}{l} F(t_1, x_1, \dot x_1) - [\dot \phi_1(t_1) - \dot x_1] F_{\dot x} (t_1, x_1, \dot x_1) = 0, \\ F(t_2, x_2, \dot x_2) - [\dot \phi_2(t_2) - \dot x_2] F_{\dot x} (t_2, x_2, \dot x_2) = 0. \end{array} \right\} $$

(2)

If the equations of the curves along which the left- and right-hand end-points are displaced are given in implicit form, $\omega_1(t,x) = 0$ and $\omega_2(t,x) = 0$, then the transversality condition (1) can be written in the form

$$ \left. \begin{array}{ll} \dfrac{F - \dot x F_{\dot x}}{w_{1t}} = \dfrac{F{\dot x}}{\omega_{1x}} & \text{at the left-hand end-point,} \\ \dfrac{F - \dot x F_{\dot x}}{w_{2t}} = \dfrac{F{\dot x}}{\omega_{2x}} & \text{at the right-hand end-point} \end{array} \right\} $$

(3)

If there are no constraints on one of the end-points, then at this end-point, by virtue of the independence of the respective tangent differentials $dt$ and $dx$, the transversality condition takes the form

$$ F = 0, \qquad F_{\dot x} = 0. $$

(4)

The relations (2), (3), (4) are called transversality conditions.

Below, the transversality condition is given in the more general case of the variational problem for a conditional extremum. Consider the Bolza problem, i.e. the problem of minimizing a functional

$$ \begin{gathered} J(x) = \int_{t_1}^{t_2} f(t,x,\dot x) dt + g(t_1, x(t_1), t_2, x(t_2)), \\ f:\R \times \R^n \times \R^n \to \R, \qquad g:\R \times \R^n \times \R \times \R^n \to \R, \end{gathered} $$

(5)

in the presence of differential constraints of equality type,

$$ \phi(t,x,\dot x) = 0, \qquad \phi: \R\times \R^n\times \R^n\to\R^m, \quad m<n, $$

(6)

and boundary conditions

$$ \begin{gathered} \psi(t_1, x(t_1), t_2, x(t_2)) = 0,\\ \psi: \R \times \R^n\times\R\times\R^n\to\R^p, \qquad p\le 2n+2. \end{gathered} $$

(7)

When $p < 2n+2$ in this problem, the end-points $(t_1, x_1^1, \ldots, x_1^n)$ and $(t_2, x_2^1, \ldots, x_2^n)$ of the extremal are not fixed, but can be displaced along given hypersurfaces $\psi_\mu = 0$, $\mu = 1, \ldots, p$.

In accordance with the transversality condition, there exist constants (Lagrange multipliers) $e_\mu$, $\mu=1,\ldots,p$, as well as multipliers $\lambda_0$ and $\lambda_i(t)$, $i=1,\ldots,m$, such that, in addition to the boundary conditions (7), the following relation holds at the end-points of the extremal:

$$ \left[ \left( F - \sum_{i=1}^n \dot x^i F_{\dot x^i}\right) \ dt + \sum_{i=1}^n F_{\dot x^i} \ dx^i \right]_1^2 + \lambda_0 \ dg + \sum_{\mu=1}^p e_\mu \ d\psi_\mu =0, $$

(8)

for any choice of tangent differentials

$$ dt_1, dx_1^i, dt_2, dx_2^i, \qquad i = 1,\ldots, n, $$

(9)

of the manifold defined by (7). In (8), $F$ denotes the expression

$$ F = F(t,x,\dot x, \lambda) = \lambda_0 f(t,x,\dot x) + \sum_{i=1}^m \lambda_i(t) \phi_i(t, x, \dot x). $$

(10)

In the majority of practical problems, the Lagrange multipliers are normalized by setting $\lambda_0=1$ (the value $\lambda_0=0$ corresponds to an abnormal case, see [1]). The multipliers $\lambda_i(t)$, $i=1,\ldots,m$, are determined together with the $x^i(t)$, $i=1,\ldots,m$, from the solution of the Euler system of differential equations

$$ F_{x^i} - \frac{d}{dt} F_{\dot x^i} = 0, \qquad i = 1,\ldots,n, $$

(11)

and $m$ equations of the form (6):

$$ \phi_i(t,x,\dot x) = 0, \qquad i=1,\dots, m. $$

The general solution of this system of $n$ second-order differential equations and $m$ first-order differential equations in the $n+m$ unknown functions $x^i(t)$, $i=1,\ldots,n$, and $\lambda_i(t)$, $i=1,\ldots,m$, depends on $2n$ arbitrary constants. In fact, if one sets

$$ \dot x^i = u_i, \qquad i = 1,\ldots,n, $$

(12)

then one obtains a system (11), (12) of $2n$ first-order differential equations and $m$ finite relations

$$ \phi_i(t,x,u) = 0, \qquad i = 1,\ldots,m. $$

(13)

Using (13), some $m$ of the functions $u_i$ can be expressed in terms of the others (under the hypothesis that the corresponding functional determinant does not vanish) and, on substituting these in (11), (12), one obtains a system of $2n$ first-order differential equations with $2n$ unknown functions, the general solution of which depends on $2n$ arbitrary constants. Along with the values $t_1$ and $t_2$, this gives $2n+2$ arbitrary constants, determining the solution of the variational problem (5)–(7). One then obtains by means of the transversality condition the correct number of equations enabling one to determine these arbitrary constants.

In problems of optimal control and in the Pontryagin maximum principle, the necessary transversality condition is written similarly to (8), only instead of

$$ F - \sum_{i=1}^n \dot x^i F_{\dot x^i} \qquad \text{and} \qquad F_{\dot x^i} $$

one has to substitute in (8) the Hamiltonian $H$, taken with the opposite sign, and the conjugate variables $\psi_i$.

The necessary transversality condition gives the missing boundary conditions for obtaining a closed boundary value problem to which the solution of the variational problem with variable end-points reduces.

References

[1] G.A. Bliss, "Lectures on the calculus of variations" , Chicago Univ. Press (1947)
[2] M.A. Lavrent'ev, L.A. Lyusternik, "A course in variational calculus" , Moscow-Leningrad (1950) (In Russian)
[a1] L. Cesari, "Optimization - Theory and applications" , Springer (1983)
[a2] L.D. Berkovitz, "Optimal control theory" , Springer (1974)
[a3] L.E. [L.E. El'sgol'ts] Elsgolc, "Calculus of variations" , Pergamon (1961) (Translated from Russian)
[a4] R.H. Rishel, "Deterministic and stochastic optimal control" , Springer (1975)
How to Cite This Entry:
Transversality condition. Encyclopedia of Mathematics. URL: http://encyclopediaofmath.org/index.php?title=Transversality_condition&oldid=14507
This article was adapted from an original article by I.B. Vapnyarskii (originator), which appeared in Encyclopedia of Mathematics - ISBN 1402006098. See original article