Can anyone give examples of mechanics problems which can be solved by Lagrange equations of the first kind, but not the second kind?


1) Let us call the equations

$$ \frac{d}{dt} \left(\frac{\partial L}{\partial \dot{q}^j}\right)-\frac{\partial L}{\partial q^j}~=~0 \qquad (s)$$ for Lagrange equations of second kind in strong sense, and let

$$ \frac{d}{dt} \left(\frac{\partial L}{\partial \dot{q}^j}\right)-\frac{\partial L}{\partial q^j}~=~Q_j\qquad (w) $$

be Lagrange equations of second kind in weak sense, where $Q_j$ are the generalized forces that don't have generalized potentials.

2) A Frictional force like

$${\bf F}_f ~=~- k{\bf v}$$

can be modeled in Lagrangian mechanics using Rayleigh's dissipation function

$${\cal F}~=~ \frac{k}{2} {\bf v}^2.$$

It will not be Lagrange equations of second kind in the strong sense $(s)$, but only in the weak sense

$$ \frac{d}{dt} \left(\frac{\partial L}{\partial \dot{q}^j}\right)-\frac{\partial L}{\partial q^j}~=~ -\frac{\partial{\cal F} }{\partial \dot{q}^j}. $$

3) Another source of examples are non-holonomic constraints. This is in particular true if one insists on the strong form $(s)$.

It is a bit more tricky if one allows the weak form $(w)$, and the $m$ non-holonomic constraints are on semi-holonomic form, say for simplicity of the form

$$ \sum_i a_i(q,t)~{\rm d}q^i + a_t(q,t)~{\rm d}t ~=~ 0, $$

where the functions $a_i(q,t)$ and $a_t(q,t)$ do not depend on the generalized velocities $\dot{q}^j$.

Then it is possible to introduce $m$ Lagrange multipliers $\lambda^\alpha$ so that the $n$ Lagrange equations become of second kind in the weak form $(w)$. However, that is not the full set of equations. Together with the $m$ constraints

$$ \sum_i a_i(q,t)~\dot{q}^i + a_t(q,t) ~=~ 0, $$

the full system will have $m+n$ equations, corresponding to $n$ $q$'s and $m$ $\lambda$'s.


  1. Herbert Goldstein, Classical Mechanics, Chapter 1 and 2.
| cite | improve this answer | |
  • $\begingroup$ Nice answer, but this seems to be all about Lagrange's equations of the second kind. $\endgroup$ – Physiks lover Nov 20 '11 at 21:42
  • $\begingroup$ Well, not according to the definition of second kind in the Springerlink that you provided in the question(v3). In the link it is explicitly stated above its eq. (4) that the constraints should be holonomic. $\endgroup$ – Qmechanic Nov 20 '11 at 22:12

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.