System Dynamics Engine in Python

This is a Python implementation of System Dynamics by Jay Forrester in Python. Additionally to the existing features coined by J. Forrester, I added new functionality with the objective of simplying the modeling process, making System Dynamics more accessible for new users in corporate environments. The objective of this simulation engine is the simulation of business models. The System Dynamics engine in based on JdHerman's implementation, although many changes have been made.

The backend code is running on a demo server 167.71.32.209:8000.

Features

• Supported modeling elements: stocks, flows, clouds, variables, data sources
• Object-oriented and JSON/GUI interface for loading SD models
• Easy formula definition
• Use external data sources directly in formulas

Usage

Install required Python packages:

• networkx
• cexprtk
• numpy
• msgfy

Import

from SystemDynamicsEngine import Engine

This demo will simulate a predator-prey dynamic system. First define the simulation time steps

delta_time = 0.1
t = np.arange(0, 40, delta_time)
eng = Engine(t)

Then define all variables used in the model

eng.variables({'a': 0.5, 'b': 0.3, 'c': 0.9})

Next, initialize the stocks

eng.stocks({'r': 10, 'p': 5})

The flows:

eng.flow('f0', start=None, end='r', f="a*r")
eng.flow('f1', start='r', end=None, f="b*r*p")
eng.flow('f2', start=None, end='p', f="b*r*p")
eng.flow('f3', start='p', end=None, f="c*p")

Finally run the simulation and optionally plot the results

eng.run()
plt.plot(t, eng.r, 'r')
plt.plot(t, eng.p, 'g')
plt.show()