Detailed Description

Abstraction for Monte Carlo moves.

A move proposes an update of the Markov chain and returns the Metropolis acceptance ratio. The framework supplies two type-erased containers:

triqs::mc_tools::move wraps a user object satisfying triqs::mc_tools::MCMove, tracks per-move proposal and acceptance counters, optional timing, and forwards the optional reject, calibrate and collect_statistics hooks.
triqs::mc_tools::move_set combines several moves with relative weights and itself satisfies the MCMove concept: on each step it draws one of its sub-moves with probability proportional to its weight, allowing composite update schemes to be assembled hierarchically.

Acceptance ratios and timing statistics are exposed for diagnostics and are aggregated across MPI ranks when results are collected.

Classes
class	triqs::mc_tools::move< MCSignType >
	Type erasure class for MC moves. More...
class	triqs::mc_tools::move_set< MCSignType >
	Set of MC moves. More...

Classes