# Principle of Least Action help me clear it up!

1. Aug 10, 2011

### Master J

I've been reading Landau and Lifgarbagez's Mechanics and have some issues I need clearing up, so I hope folk here can help :)

It is stated that the Principle of Least Action is not always valid for the entire path of a system, but only for a sufficiently small segment...what does this mean?

I interpret this as saying:

"the action will not always be minimized (altho it usually is in most cases), but it may take on a maximum value. What is important is that it is an extremum. However, if we look at a sufficiently small part of ANY path, it will always be a minimum, but the entire path may not be" ....is this correct???

2. Aug 10, 2011

### atyy

3. Aug 10, 2011

### jjustinn

The easiest-to-visualize example I've come across is geodesics ('straight lines') on the surface of a sphere; all geodesics satisfy "minimal action", but they're only guaranteed to be an actual minimum if they're "sufficiently short". For example, there are two geodesics connecting Los Angeles, CA to New York, NY on the surface of the Earth -- one going across mainland USA, and the other going around the other side of the world. Both are "minimal" in the sense of an action principle (by definition -- they're "straight lines" / geodesics), but obviously only one is a truly minimal path. Now if you were restricted to choosing paths that were shorter than one half the circumference of the Earth, you would be guaranteed to have a minimum -- due to peculiarities of a sphere.

Hopefully I got all of the facts right there, and it was clear enough to at least get the general idea across...if I failed on either front, hopefully someone will correct the record.

4. Aug 10, 2011

### mathfeel

In this example, while the continental path is a global minimum, both paths are local minimum in the functional space of all LA-NY paths: it is stationary against small variation of path and in fact has a positive second order correction. Is this what L&L had in mind?

We can all agree that geodesics are locally straight.