I understand that if a constraint equation given on a differential form is exact, that means it is also holonomic since I can find a solution. But there are other types of differential equations, like separable and linear, such that I can find an equation in the form of a holonomic constraint. Why is it that being exact is a condition, rather than just having a solution of the type $f(x,y) = 0$?

  • 1
    $\begingroup$ $f(x,y)=0$ looks like a holonomic constraint. $\endgroup$ – Qmechanic Nov 26 '17 at 11:47
  • $\begingroup$ Well, yeah... I meant that. $\endgroup$ – Ana Luiza Ferrari Nov 27 '17 at 20:10
  1. Here I would like to mention the notion of a semi-holonomic constraint $$ \sum_{j=1}^n a_j(q,t)~ \mathrm{d}q^j + a_t(q,t)~\mathrm{d}t~=~0, \tag{1'}$$ which puts an inexact differential equal to zero.

  2. It is possible to incorporate semi-holonomic constraints into Lagrange equations, cf. e.g. my Phys.SE answer here.

  3. Note that eq. (1') does not mean that we demand that the $n+1$ co-vector components $$ a_t(q,t)~= a_1(q,t)~= \ldots ~=~a_n(q,t)~=~0 \qquad\qquad (\longleftarrow \text{Wrong!} ) \tag{2'}$$ of a co-vector (1') should be zero. Perhaps this potential misunderstanding (2') is secretly the core of OP's question?

  4. Rather eq. (1') means that $$ \sum_{j=1}^n a_j(q,t)~ \dot{q}^j + a_t(q,t)~=~0. \tag{3'}$$

  5. In particular, a holonomic constraint $$f(q,t)~=~0 \tag{0}$$ can be put on the above semi-holonomic form $$ \mathrm{d}f~=~\sum_{j=1}^n \frac{\partial f}{\partial q^j}~ \mathrm{d}q^j + \frac{\partial f}{\partial t}~\mathrm{d}t~=~0 .\tag{1}$$ Explicitly, eq. (1) means that the total time derivative $$ \frac{df}{dt}~\equiv~ \sum_{j=1}^n \frac{\partial f}{\partial q^j}~ \dot{q}^j + \frac{\partial f}{\partial t}~=~0\tag{3}$$ is zero.


  1. H. Goldstein, Classical Mechanics, Section 2.4.

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.