Marmote’s documentation
Table of Contents
Marmote
It consists in a library of objects:
allowing to construct custom Markov Chain models, analyze their structure and solve them
allowing to construct custom Markov Decision Process models, solve them and analyze the solution
representing well-known Markovian models from the literature.
This library comes with two API: one for C++ development, one for Python development (that encapsulates C++ objects).
You can learn how to
build a consistent development environment for contributing to marmote, using conda
program your models with the C++ API
program your models with the Python API
Latest Versions
Version 1.2 released on 2024-05-31 (works with python 3.10 and 3.11)
Version 1.1.1 released on 2024-05-21 (deprecated)
Version 0.3.0 released on 2023-10-12 (deprecated)
Licence
This software is distributed under the terms of the GNU General Public License (GPL) version 3. By using, modifying or distributing this software, you agree to the terms and conditions of the GPL.
Features
Predefined Spaces (Interval, Box, Composite Set, Simplex, Binary Sequence)
Easy handling of Transition Structures
Support sparse matrices
Markov Chain Features
Graph Analysis of transitions
Discrete and Continuous Time Chains
Predefined Chains (Birth and Death, Random Walk, Poisson Process, MMPP)
Usual solving methods (Power Method)
Advanced solving methods (RLGL)
Markov Decision Process Features
Criteria implemented are: infinite horizon discounted, infinite horizon average, infinite horizon total, finite horizon discounted
Solving Methods are Value Iteration and Policy Iteration Modified
Structural properties analysis
For more details about features, you have access to the c++ Documentation including the user/reference manual and the Python Documentation with the detailed API documentations.
Documentation creation date: May 22 2025 at 18:49 +0200