ASDM is a python library that enables its users to create and simulate System Dynamics (SD) models. It can also simulate SD models saved in the XMILE format, including advanced features such as arrays and conveyors. The support is being continuously improved.
In the library:
-
asdm/asdm.pyconsists of the main functionalities, including the lexer, parser, and interpreter. -
asdm/utilities.pyprovies a data visualisation tool. -
asdm/Inferenceconsists of tools for model calibration.
The library is developed and used with Python 3.11. However, it should also work with other versions. If you encounter a problem and believe it is related to python version, please open an issue or contact me.
The library is developed and used on macOS. Some tests have also been done on Windows and Ubuntu Linux, but they are not comprehensive. If you encounter a problem and believe it is related to the OS, please open an issue or contact me.
We recommend that you create a new environment to use the library, although this is not always necessary. For example, if you use anaconda, this can be done by:
conda create --name asdm
To clone this repository to your local environment, please ensure that git is installed in your system, then use the following command:
git clone https://github.com/wzh1895/ASDM.git
ASDM relies on a number of other python libraries as dependencies. To install them, use the following commands:
cd asdm
conda install --file requirements.txt -c conda-forge
At the root folder of this repository, execute the following code in the interactive Python environment or as a part of a script:
from asdm.asdm import sdmodel
pip install asdm
ASDM and its required dependencies will be automatically installed.
At any path, execute the following code in the interactive Python environment or as a part of a script:
from ASDM.asdm import sdmodel
Note: there is a difference in code between the above two ways of imports. We are working on making them consistent in future updates.
Please refer to Documentation for the commonly used functions.
We also use Jupyter Notebooks to provide demoes of ASDM's functionalities.
- Creating an SD model from scratch
- Adding stocks, flows, auxiliaries
- Support for nonlinear functions (MIN, MAX, etc.)
- Support for stochastic functions (random binomial trial, etc.)
- Running simulations
- Exporting and examing simulation outcomes
- Displaying simulation outcomes as graph
We will add more tutorial notebooks.
You are welcomed to share your own tutorial notebooks through a pull request. When sharing notebooks, please make sure it does not contain sensitive data.
ASDM is made public under the MIT licence.
Wang Zhao main author
PhD candidate & research assistant at University of Strathclyde, UK
Wang has given multiple talks on ASDM at different gatherings and conferences of modellers, operational researchers, and healthcare experts. This is the YouTube link to one of the talks.
Wang can be reached at wang.zhao@strath.ac.uk.
Matt Stammers contributor
Consultant Gastroenterologist & open-source software developer at University Hospital Southampton, UK
Matt created a webapp using streamlit to allow users to interact with a simulation dashboard in their web browsers or on smartphones, such as in this demo. The simulation of the SD model in the backend is powered by ASDM. This is a part of an initiative Really Useful Models.
Matt's GitHub Homepage