- #1
espen180
- 834
- 2
When going from the Lagrangian to the Hamtiltonian, we define
[tex]p_i=\frac{\partial L}{\partial \dot{q}_i}[/tex]
as the independent variables in place of [tex]\dot{q}[/tex]. This change of variables is possible if and only if the Hessian matrix
[tex]\frac{\partial^2L}{\partial \dot{q}_i\partial\dot{q}_j}[/tex]
is non-singular. The Hamiltonian is then defined by performing a Legendre transformation on the Lagrangian:
[tex]H(q,p,t)\equiv \sum_{i=1}^n \dot{q}_i p_i - L(q,\dot{q},t)[/tex].
Since the first Hamilton's equation is
[tex]\dot{q}_i=\frac{\partial H}{\partial p_i}[/tex]
, this implies that we can perform the inverse change of coordinates if and only if the Hessian matrix
[tex]\frac{\partial^2 H}{\partial p_i \partial p_j}[/tex]
is non-singluar, and we would re-obtain [tex]L[/tex] as
[tex]L(q,\dot{q},t)\equiv \sum_{i=1}^n p_i \dot{q}_i - H(q,p,t)[/tex]
Now for my questions.
1) Does there exist a physical system with a Lagrangian for which [tex]\frac{\partial^2L}{\partial \dot{q}_i\partial\dot{q}_j}[/tex] is singular, or a physical system with a Hamiltonian such that [tex]\frac{\partial^2 H}{\partial p_i \partial p_j}[/tex] is singular?
2) Does the non-singularity of [tex]\frac{\partial^2L}{\partial \dot{q}_i\partial\dot{q}_j}[/tex] immediately imply the non-singularity of [tex]\frac{\partial^2 H}{\partial p_i \partial p_j}[/tex]? If so, how? How about the opposite case (singular implies singular)? Thanks for any help.
[tex]p_i=\frac{\partial L}{\partial \dot{q}_i}[/tex]
as the independent variables in place of [tex]\dot{q}[/tex]. This change of variables is possible if and only if the Hessian matrix
[tex]\frac{\partial^2L}{\partial \dot{q}_i\partial\dot{q}_j}[/tex]
is non-singular. The Hamiltonian is then defined by performing a Legendre transformation on the Lagrangian:
[tex]H(q,p,t)\equiv \sum_{i=1}^n \dot{q}_i p_i - L(q,\dot{q},t)[/tex].
Since the first Hamilton's equation is
[tex]\dot{q}_i=\frac{\partial H}{\partial p_i}[/tex]
, this implies that we can perform the inverse change of coordinates if and only if the Hessian matrix
[tex]\frac{\partial^2 H}{\partial p_i \partial p_j}[/tex]
is non-singluar, and we would re-obtain [tex]L[/tex] as
[tex]L(q,\dot{q},t)\equiv \sum_{i=1}^n p_i \dot{q}_i - H(q,p,t)[/tex]
Now for my questions.
1) Does there exist a physical system with a Lagrangian for which [tex]\frac{\partial^2L}{\partial \dot{q}_i\partial\dot{q}_j}[/tex] is singular, or a physical system with a Hamiltonian such that [tex]\frac{\partial^2 H}{\partial p_i \partial p_j}[/tex] is singular?
2) Does the non-singularity of [tex]\frac{\partial^2L}{\partial \dot{q}_i\partial\dot{q}_j}[/tex] immediately imply the non-singularity of [tex]\frac{\partial^2 H}{\partial p_i \partial p_j}[/tex]? If so, how? How about the opposite case (singular implies singular)? Thanks for any help.