8.5 Time Evolution of the Distribution Function

The Liouville equation for a system subject to a time dependent external field is given by

(8.20)

where we have defined the time dependent f-Liouvillean, iL(t). This equation tells you that if you sit at a fixed point in phase space denoted by the dummy variable Γ, the density of phase points near Γ, changes with time in accord with (8.20). In the derivation of this equation we related the partial derivative of f(t) to various fluxes in phase space at the same value of the explicit time.

We define the distribution function propagator UR(0,t) which advances the time dependence of the distribution function from 0 to t, by

(8.21)

In this equation UR (0,t) is the adjoint of UR(0,t). It is therefore closely related to UL(0,t) except that the Liouvilleans appearing in equation (8.3.7) are replaced by their adjoints iL(si). Combining equation (8.5.2) with the Liouville equation (8.5.1) we find that UR (0,t) satisfies the following equation of motion

(8.22)

The formal solution to this operator equation is

(8.23)

In distinction to the propagator for phase variables, the integration limits imply that t > s1 > s2 > .... > sn , so that the f-Liouvilleans are left time ordered. The time arguments increase as we go from the right to the left. This is opposite to the time ordering in the p-propagator UR(0,t) but the definition of UR (0,t) is consistent with the definition of UL(0,t).

For the f-propagator U†(0,t), the usual associative law is satisfied as the time arguments are ordered right to left,

(8.24)

This equation can be verified directly using similar arguments to those used in §8.4.