In physics, and more specifically in Hamiltonian mechanics, a generating function is, loosely, a function whose partial derivatives generate the differential equations that determine a system's dynamics. Common examples are the partition function of statistical mechanics, the Hamiltonian, and the function which acts as a bridge between two sets of canonical variables when performing a canonical transformation.
There are four basic generating functions, summarized by the following table:[1]
Sometimes a given Hamiltonian can be turned into one that looks like the harmonic oscillator Hamiltonian, which is
H = a P 2 + b Q 2 . {\displaystyle H=aP^{2}+bQ^{2}.}
For example, with the Hamiltonian
H = 1 2 q 2 + p 2 q 4 2 , {\displaystyle H={\frac {1}{2q^{2}}}+{\frac {p^{2}q^{4}}{2}},}
where p is the generalized momentum and q is the generalized coordinate, a good canonical transformation to choose would be
This turns the Hamiltonian into
H = Q 2 2 + P 2 2 , {\displaystyle H={\frac {Q^{2}}{2}}+{\frac {P^{2}}{2}},}
which is in the form of the harmonic oscillator Hamiltonian.
The generating function F for this transformation is of the third kind,
F = F 3 ( p , Q ) . {\displaystyle F=F_{3}(p,Q).}
To find F explicitly, use the equation for its derivative from the table above,
P = − ∂ F 3 ∂ Q , {\displaystyle P=-{\frac {\partial F_{3}}{\partial Q}},}
and substitute the expression for P from equation (1), expressed in terms of p and Q:
p Q 2 = − ∂ F 3 ∂ Q {\displaystyle {\frac {p}{Q^{2}}}=-{\frac {\partial F_{3}}{\partial Q}}}
Integrating this with respect to Q results in an equation for the generating function of the transformation given by equation (1):
F 3 ( p , Q ) = p Q {\displaystyle F_{3}(p,Q)={\frac {p}{Q}}}
To confirm that this is the correct generating function, verify that it matches (1):
q = − ∂ F 3 ∂ p = − 1 Q {\displaystyle q=-{\frac {\partial F_{3}}{\partial p}}={\frac {-1}{Q}}}