My Lagrangian equation is $$L = \dfrac{1}{2}m\dot{q}^{2} \tag{1},$$ where $q=(x,y)$. Performing the Legendre transformation I get the Hamiltonian equation, \begin{equation} H(p,q) =p\dot{q}-\dfrac{1}{2}m\dot{q}^{2}\tag{2}. \end{equation}

Now from the Lagrange equations it follows obviously that $p=m\dot{q}$ which gives that, $H(p,q)= \dfrac{1}{2}m\dot{q}^{2}$. If I put this hamiltonian equation in cartesian coordinates $(x,y)$ then my Hamiltonian will become, \begin{equation} H(x,y) = \dfrac{1}{2}m(\dot{x}^{2}+\dot{y}^{2}).\tag{3} \end{equation}

My question is as follows:

Are my $dH$ and my Hamiltonian vector field $X_{H}$ defined as follows?

\begin{equation} dH(x,y) = \dfrac{\partial H}{\partial x}dx + \dfrac{\partial H}{\partial y}dy \\ X_{H} = \dfrac{\partial H}{\partial x }\dfrac{\partial}{y} -\dfrac{\partial H}{\partial y}\dfrac{\partial}{\partial x}. \end{equation}

or as this?

\begin{equation} dH(p_{x},p_{y},x,y) = \dfrac{\partial H}{\partial x}dx + \dfrac{\partial H}{\partial y}dy +\dfrac{\partial H}{\partial p_{x}}dp_{x} + \dfrac{\partial H}{\partial p_{y}}dp_{y} \\ X_{H} = \dfrac{\partial H}{\partial x }\dfrac{\partial}{p_{x}} -\dfrac{\partial H}{\partial p_{x}}\dfrac{\partial}{\partial x}+\dfrac{\partial H}{\partial y }\dfrac{\partial}{p_{y}} -\dfrac{\partial H}{\partial p_{y}}\dfrac{\partial}{\partial y}. \end{equation}

  1. Note that OP's function (3) in it current form (v3) is the Lagrangian energy function $$h(q,\dot{q},t)~:=~\sum_{i=1}^n\dot{q}^i\frac{\partial L(q,\dot{q},t)}{\partial \dot{q}^i}-L(q,\dot{q},t);$$ not the Hamiltonian $H(q,p,t)$ as OP claims. The two functions have same value, but depend on different arguments. In fact, the Lagrangian and the Hamiltonian formalism are connected via a Legendre transformation between the variables $\dot{q}$ and $p$.

  2. The Hamiltonian vector field $X_H\in\Gamma(T^{\ast}M)$ [with the Hamiltonian$^1$ $H(q,p)$ as generator] is a vector field on the cotangent bundle/phase space $T^{\ast}M$; not on the configuration space $M$.


$^1$ Assume for simplicity no explicit time dependence.


In simplified words said the purpose of the Hamiltonian formalism is to eliminate the velocities $\dot{q}$ in favour of the momenta $p$. When you combine $p=m\dot{q}$ with $H=\frac{1}{2}m \dot{q}^2$, the velocity $\dot{q}$ should be eliminated in favour of $p$. If this is done we get ( in 2 dimensions):

$H = \frac{p^2}{2m} \equiv \frac{p_1^2}{2m} + \frac{p_2^2}{2m}$

which only depends on $p$. In that case the components of the Hamiltonian vector field are

$X_H = \left( \frac{\partial H}{\partial p_i}, -\frac{\partial H}{\partial q_i}\right) = (p_i, 0)$

if you like them expanded in a base of tangential vectors in 2 dimensions.

$X_H = p_1 \frac{\partial }{\partial q_1} + p_2 \frac{\partial }{\partial q_2} + 0 \frac{\partial }{\partial p_1} + 0 \frac{\partial }{\partial p_2} = p_1 \frac{\partial }{\partial q_1} + p_2 \frac{\partial }{\partial q_2}$

The symplectic 2-form in 2 dimensions is independent of the dependency of $H$ on $q$ and $p$: $dq_1\wedge dp_1 + dq_2\wedge dp_2$. Because the phase space of a 2-dimensional problem is always the same. Of course the curves $H(q,p)=const$ (here $H(p)=const$) in the phase space might be different for each different 2-dim. problem.


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.