Simple Monte Carlo Library
Copyright 2008 James Ravn
james.ravn@gmail.com
Introduction
------------
The goal is to provide a monte carlo library that is fast, portable,
and simple. The only thing required by the user is to create their
simulation.
Features
--------
* Bootstrap algorithm
* Error calculation using bootstrap
* High quality pseudo random number generator
* Distribution generators
* No required libraries apart from the C90 standard library
Where To Get It
---------------
The latest version can be cloned using git:
git clone git://git.jsravn.com/mc.git
Please send bug reports and patches to james.ravn@gmail.com.
Testing & Compiling
-------------------
Run 'make test' to compile and run all the tests. Take a look at the
unit tests (files ending with _test.c) and Makefile to get an idea of
how to compile your own project.
Using
-----
The details of using the library are provided in the header
files. Mc.h is a good starting point.
The basic steps are:
1. Write a function that executes a single iteration of your
simulation. Have it return a float. For example, return 1.0 if it is
successful, 0.0 if it is not.
2. Initialize the bootstrap error algorithm (see mc_init_bs).
3. Pass in your function to mc_simulate_many. The results can be
accessed using the functions mc_mu(), mc_err(), and mc_var().
Portability
-----------
It is written in ANSI C, requiring only the long long C99 extension for
the provided psuedo random number generator. If needed it is easy to
write a C90 compatible psuedo random number generator. Just create your
implementation of mc_prng.h.
Future ideas
------------
* Create some example simulations
* Create interfaces to work w/ Python/Ruby