Monte Carlo Methods for Applied System Scientists
The name Monte Carlo Method generally applies to various applications of pseudo random number generators to get information about mathematical expressions. I like to focus on uncertainty examination of mathematical models, which consist of a few analytic functions or numeric solutions to differential equations. That is, what i need. :-) This list is by no means complete, main purpose is to share my experiences with this special subject and therefor i will include my personal comments.First search using the internet was done 12/13.6.1997. The second after reading the NCRP Commentary No.14 (see below) after the 5th of september 1997.
Introductions and Linklists
- History of Monte
Carlo Method
Short historic text about the
first people to discover that method.
- Introduction to
Monte Carlo Methods A full featured elektronic
book, which introduces in the theory, starting with basic probability
and statistics.
- Random number generators --
The pLab Project Home Page (at Uni Salzburg)
Random Generators
are the heart of any monte carlo implementation. They must be
choosen with care and tested thoroughly.
- The World-Wide Web
Virtual Library: Random Numbers and Monte
Carlo methods
The pLab research group maintains this page. Contains links to some
software products.
Free Software
-
MCSim
Monte Carlo Simulation Software
Package written in C, which can solve systems of ODEs numerical,
run monte carlo analysis on numerical or analytical models and can
do Markov Chain Monte Carlo simulations in batch mode.
It is Free Software covered
by the GNU General Public License and should run on any platform for
which a C-compiler, like GNU`s gcc is available.
You can also get it from an GNU Software Archive.
This seems so be quite usable for my tasks so far.
Write your own?
Many numeric libraries do have support for various random generators and could be used to write own monte carlo programs. The same holds for general mathematic programs, like Mathematica, Maple or MuPAD.Proprietary, Commercial Software
Some proprietary commercial products are available, which often offer the monte carlo methods together with spreadsheets or decision support systems. It is called usually "RiskAnalysis" in this context. (;
- Cristal Ball from
Decisioneering,Inc. An
add-on to Excel or Lotus 1-2-3 spreadsheets.
Stefan Schwartz is using this experimental for his
work in the project:
Validierung von Umweltexpositionsmodellen
und in Modellen verwendeten Parametern
(in german)
- DynRisk
from TerraMar Information Systems
Includes links to other risk related sites.
- DecisionPro
from Vanguard
- @Risk
from Palisade Corporation
- Analytica
Succesor to DEMOS from Lumina Decision Systems,
Inc.
Support Statistic Software
If you want to apply monte carlo methods, you do need statistical support functionality to get the input distributions right and to interpret the results. The range of software is broad here. Some packages have several statistic functions integrated, others rely on external programs. For completeness i link only to statist,
which is small tool for
standard statistical tasks. It is free software and I use it.
[This subject really demands a separate page.]
Other; mentioned in NCRP Commentary No. 14
The NCRP Commentary No. 14
titled "A
Guide For Uncertainty Analysis in Dose
and Risk Assessments Related to Environmental Contamination" from may 1996
mentiones a few other Monta Carlo Simulation tools apart from
Cristal Ball and @Risk i mentioned above. Here are the references i
found on the web, it all looks quite old:
- MOUSE
(Modular Oriented Uncertainty SystEm)
-
TAM3Demonstrates Monte Carlo Sensitivity and
Uncertainty Analysis.
- PRISM (no info found)
-
PSR-0275: ORMONTE,Uncertainty Analysis for
User-Developed System Models Can be linked to
models implemented in FORTRAN.
- DEMOS (webpage vanished in 1998) has been further developed into Analytica (see above)
-
GENII-S Environmental Radiation Dosimetry
Software System
- RiskQ (no info found)
- SUSA-PC (no info found)
E-Mail: Bernhard.Reiter@usf.uni-osnabrueck.de
[ Bernhards Heimatseite(»German) ] [ Diplomarbeit Bernhard Reiter(»German) ]