Overview

This package provides a collection of types and routines for realizing path-integral Monte Carlo procedures in occupation number representation. In particular, one samples many-particle trajectories in imaginary time which are represented by the type Configuration{T}.

It is given the set of orbitals occupied at $\tau = 0$, and a collection of one- and two-particle excitations and their respective imaginary times. The type used for representing single-particle basis states is arbitrary and corresponds to the type variable T.

One generally needs two kinds of functions:

• Updates, which propose changes to a Configuration, thus providing the building blocks of the Markov chain.
• Estimators, which give the contribution of a particular Configuration to an expectation value.

The Monte Carlo process is carried out by the function sweep!, which generates the Markov chain from which the expectation values are calculated.

A many-particle model is specified by the one- and two-particle matrix elements $\epsilon_{ij}$ and $w_{ijkl}$ with respect to the chosen basis. These can be used to calculate the weight changes for the acceptance probability.