Muller Documentation

Implements a suite of sampling algorithms to stochastically draw points from Muller's potential. Provides a framework to perform a comparative analysis of samplers for Monte Carlo integration.

Licensed under Creative Commons BY-NC-SA 3.0.

Questions or comments? Contact Joseph Schlecht.

usage: muller OPTIONS
where OPTIONS is one or more of the following:
 -h, --help              Prints program usage.
 -v, --version           Prints program version.
     --seed=ARG          Random seed.
     --iterations=ARG    Number of sampling iterations.
     --hybrid-leap-iter=ARG  
                         Number of hybrid Monte Carlo leapfrog iterations.
     --start-x=ARG       Starting x-coord position.
     --start-y=ARG       Starting y-coord position.
     --delta-t=ARG       Stochastic dynamics integration step-size for Langevin
                         Monte Carlo.
     --alpha=ARG         Stochastic dynamics step-size for Langevin Monte Carlo.
     --kick=ARG          Stochastic dynamics frequency to kick the particle by
                         resetting the momenta. Zero disables any kicking. Must
                         be > 0.
     --x-sigma=ARG       Gaussian sigma for MH proposal of x. Must be > 0.
     --x-lim=ARG         Limits over the range of x. Has format min,max
     --y-sigma=ARG       Gaussian sigma for MH proposal of y. Must be > 0.
     --y-lim=ARG         Limits over the range of y. Has format min,max.
     --sampler=ARG       Type of sampler to use. Must be one of {mh, lange,
                         hyper, stoch, hybrid}.
     --hyper-bias=ARG    Bias parameter for hyper sampling.
     --hyper-scale=ARG   Scale parameter for hyper sampling.