EMBER (Emergence and Macroscopic Behavioural ExtRaction is a generalised analysis tool for
numerical continuation and bifurcation analysis of stochastic systems or those without a closed
form solution or macroscopic equation of state. The code is based on equation-free techniques to
side step the need for these equations and thus we can perform continuation, bifurcation analysis
and extract statistical descriptors of a system or simulation model. EMBER also contains a method
for determination of necessary equation-free parameters based on the dynamics of the system and
a robust continuation method to coping with highly noisy systems. The algorithms behind the
EMBER code are introduced in detail in this paper [1].
EMBER is written in java for compatibility with Windows, MacOS and Linux operating system,
object-orientated design (OOD) and also for the flexibility with API to external models. The OOD
enables aspects of the code to be encapsulated to increase code usability and upgrades. For instance,
the bifurcation detection is based on the change in sign of the determinate of the Jacobian, a simple
but un-robust method, which is defined in an isolated module of the code and could be substituted
for another method such as GMRES without affecting the rest of the code. Another example is
the API which is the only part of the code that interacts with the external model, therefore any
system can be analysed with this code providing a API can be written. Currently this program will
analyse java based simulations and agent-based models written in Netlogo, further developments
will extend this to other systems and languages.
Keywords | bifurcation analysis, stochastic systems, SDE, equation-free method |
Model | |
Software Type | |
Language | |
Platform | |
Availability | |
Contact Person | |
References to Papers | Spencer Angus Thomas, Anne C. Skeldon, and David B. J. Lloyd. Equation-free analysis of
agent-based models and systematic parameter determination. Physica A, 1 (1) May:142, 2016.
|