Lagrangian's formalism and Fermat's principle

Based on Lagrangian's formalism one can write Eq.(2.9) as
\begin{displaymath}
\tau = \displaystyle{ \int\limits_{A}^{B}} {\cal L} \; {ds}  ,
\end{displaymath} (2.10)

where ${\cal L}$ represents the system's Lagrangian:
\begin{displaymath}
{\cal L} = \frac{1}{c}  .
\end{displaymath} (2.11)

Within the context of the formalism ${\cal L}$ is supposed to be a function of coordinates and generalized velocities[5]:
\begin{displaymath}
{\cal L}(x,y,z,\dot{x},\dot{y},\dot{z}) = \frac{1}{c}\sqrt{ \dot{x}^2 + \dot{y}^2 + \dot{z}^2 }  ,
\end{displaymath} (2.12)

where

\begin{displaymath}
\dot{x} = \frac{dx}{ds}  , \hskip5mm
\dot{y} = \frac{dy}{ds}  , \hskip5mm
\dot{z} = \frac{dz}{ds}  .
\end{displaymath}

In this way, the perturbation of the travel time corresponds to

\begin{displaymath}
\delta \tau = \displaystyle{ \int\limits_{A}^{B}} \left\{
\...
...L}}}{\partial{\dot{z}}}\delta \dot{z} \right]
\right\} ds =
\end{displaymath}


\begin{displaymath}=
\underbrace{\left. \frac{\partial{{\cal L}}}{\partial{\dot...
... L}}}{\partial{\dot{z}}}\delta z \right\vert _{A}^{B}}_{=0} +
\end{displaymath}


\begin{displaymath}+
\displaystyle{ \int\limits_{A}^{B}}
\left\{
\left[ \fra...
...partial{\dot{z}}} \right) \right] \delta z
\right\} ds  .
\end{displaymath}

According to Fermat's principle

\begin{displaymath}\delta\tau = 0  , \end{displaymath}

which implies that
\begin{displaymath}
\begin{array}{ccccc}
\displaystyle{ \frac{d}{ds}\left( \frac...
...frac{\partial{{\cal L}}}{\partial{z}}} & = & 0  .
\end{array}\end{displaymath} (2.13)

On the other side, using Eq.(2.12) one can obtain that

\begin{displaymath}
\frac{\partial{{\cal L}}}{\partial{\dot{x}}} = \frac{\dot{x}...
...}} {\cal L} = {\cal L}\dot{x} =
\frac{1}{c}\frac{dx}{ds}  ,
\end{displaymath}


\begin{displaymath}
\frac{\partial{{\cal L}}}{\partial{\dot{y}}} = \frac{\dot{y}...
...}} {\cal L} = {\cal L}\dot{y} =
\frac{1}{c}\frac{dy}{ds}  ,
\end{displaymath}


\begin{displaymath}
\frac{\partial{{\cal L}}}{\partial{\dot{z}}} = \frac{\dot{z}...
...}} {\cal L} = {\cal L}\dot{z} =
\frac{1}{c}\frac{dz}{ds}  .
\end{displaymath}

Therefore, one can conclude that the solution of the Eikonal equation requires the solution of the following system of equations:

\begin{displaymath}
\frac{d}{ds}\left( \frac{1}{c}\frac{dx}{ds} \right) = \frac{...
...partial{}}{\partial{y}}\left( \frac{1}{c} \right) \hskip5mm ,
\end{displaymath}


\begin{displaymath}
\frac{d}{ds}\left( \frac{1}{c}\frac{dz}{ds} \right) = \frac{\partial{}}{\partial{z}}\left( \frac{1}{c} \right)  .
\end{displaymath} (2.14)

Orlando Camargo Rodríguez 2012-06-21