By George, it's Hamiltonian!

In the following problems you will show that the phase space vector field
F(x,v)=(dx/dt, dv/dt)=(v,-grad f(x))
for this system actually is Hamiltonian.

Question #3:

In the previous lab you showed that gradient vector fields have the property that work done in the field in getting between two points is independent of the path you take. Show that F(x,v) is not a gradient vector field by showing that it does not have this path-independent property. To do this, find two paths between points (0,0) and (0,3) such that the line integrals for work along each have different values.

One way to do this is to find parametrizations for your two curves and evaluate the line integrals

explicitly. We've provided some hints to do this with Maple.

Alternatively, you can give a proof without having to calculate integrals explicitly by carefully choosing the two paths based on how the geometry of the curves relative to the vector field affects the work integrals.

If you use the first method, show all your calculations and parametrizations. If you use the second method, give the geometric reasoning for choosing the curves you did. In either case, sketch your paths on the vector field plot.


Question #4:

There are plenty of vector fields that are neither gradient nor Hamiltonian, so your above results are not enough to show that F is Hamiltonian.

Question #5:

One feature of Hamiltonian systems is that the trajectories of the vector field lie on the level sets of the Hamiltonian H(x,v).
Next: Hamiltonian systems and energy
Previous: One-Dimensional Gradient Motion
Up: Introduction
Bob Thurman;thurman@geom.umn.edu>

Last modified: Tue Apr 2 13:48:52 1996