## Retrospect on statistical mechanics
Gibbs ensemble, or just say ensemble: do not focus on specific microscopic state, but a collection of system with:
- identical composition
- identical in macroscopic condition ($V, T, P$, etc.)
- exists in different state
A Gibbs ensemble is represented by a distribution of representative points in [[Potential, force, momentum, velocity#^phase-space|phase space]] $\Gamma$, and may be described as a ***distribution function*** $\rho(p,q,t)$.
>[!note]
>Here I think Gibbs ensemble is just a generalized concept, a "big" system called ensemble, and described as $\rho(p,q,t)$ with these generalized condition.
>i.e., an ensemble being described by a distribution function.
And one may define **ensemble average** as
$\langle O\rangle = \frac{\int \mathrm{d}^{3N}p \mathrm{d}^{3N}q O(p,q)\rho(p,q,t)}{\int \mathrm{d}^{3N}p \mathrm{d}^{3N}q \rho(p,q,t)}$ ^23b0cd
For dilute ideal gas, most probable velocity is
$\bar{v} = \sqrt{\frac{2kT}{m}}$
and (corresponding) energy is $\frac{3NkT}{2}$.
>[!note]
>the velocity given above is at each direction, being the same as $\frac{1}{2}mv^2 = kT$, and the energy is just 3D case of this result.