i want to arrive to hamilton-jacobi equation using the riemannian geometry.

So let $\textbf{X}\in \mathfrak{X}(M)$, where $M$ is Riemannian manifold whose metric is $g:\textbf{T}M \times \textbf{T}M \longrightarrow \mathbb{R}$. On the other hand, let suppose that $\textbf{X}=grad f$.

So $g(\textbf{X},\textbf{X})=g(g^{ij}\frac{\partial f}{\partial x^{i}}\frac{\partial }{\partial x^{j}},g^{\mu \nu}\frac{\partial f}{\partial x^{\mu}}\frac{\partial }{\partial x^{\nu}})=g^{ij}g^{\mu\nu}\frac{\partial f}{\partial x^{i}}\frac{\partial f}{\partial x^{\mu}}g(\frac{\partial }{\partial x^{j}},\frac{\partial }{\partial x^{\nu}})=g^{ij}g_{j\nu}g^{\nu\mu}\frac{\partial f}{\partial x^{i}}\frac{\partial f}{\partial x^{\mu}}$.

Finally, we get that: $g(\textbf{X},\textbf{X})=g^{i\mu}\frac{\partial f}{\partial x^{i}}\frac{\partial f}{\partial x^{\mu}}$.

Until here it is more or less the things that i know . The function that i defined before it is a lagrangian of free particle $L=\frac{1}{2}g_{ij}\dot x^{i}\dot x^{j}$, so we now that the curves of this lagrangian $x^{k}(t)$ are geodesic that is $\nabla_{\dot x^{k}}\dot x^{k}=0$. I do not know to prove with this information that $g(\textbf{X},\textbf{X})$ is a constant, where $\textbf{X}$ is the vector field generated by $L$

Finallly, i have another question of Jacobi's Principle. Let $H\in C^{\infty}(\textbf{T}^{*}M) $ be a hamiltonian function of the form $H=\frac{1}{2}g^{ij}p_{i}p_{j}+V$ where $g$ is the metric of the riemaniann manifold and $V$ is the potential function $V:M\longrightarrow \mathbb{R}$. The jacobi's principle said that the curves in $M$ are geodesic of metric $(E-V)g_{ij}dx^{i}\otimes dx^{j}$ and the new hamiltonian is $h=\frac{1}{2}(\frac{g^{ij}}{E-V})p_{i}p_{j}$. Again, the solution of the curves here are geodesic, but some of the properties of the geodisc are that the tangent curves are paralel to the curve and the aceleration is 0 so, if it is a potential, there is an aceleration. My questions are this geodeisc lies in a new manifold $M_{2}?$ If it is like this how can arrive to the original manifold?

Thanks a lot!!!!!


In a non-relativistic context the metric becomes like a position-dependent mass-tensor for the kinetic energy [1] $$ T~=~\frac{1}{2}M_{jk}(q)\dot{q}^j\dot{q}^k, \qquad M_{jk}(q)~\equiv~g_{jk}(q), \tag{8.83}$$ where $$(ds)^2~=~g_{jk}(q)~\mathrm{d}q^j ~d\mathrm{d}q^k\tag{8.86}$$ so that $$ T~=~\frac{1}{2}\left(\frac{ds}{dt}\right)^2.\tag{8.87}$$ The abbreviated action therefore becomes $$ \int p_k\dot{q}^k ~\mathrm{d}t~=~\int 2T ~\mathrm{d}t~=~\int \sqrt{2T} ~\mathrm{d}s ~=~\int \sqrt{2(E-V(q))} ~\mathrm{d}s.\tag{8.89}$$ The latter is the sought-for Jacobi's formulation.


  1. Herbert Goldstein, Classical Mechanics, Section 8.6.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.