Back to CSCSSoftware Home
Back to TFACES Home
TFACES
Under construction, Spring 2010
Summary
A large part of any computational modeling project involves analyzing and displaying the data generated, to help the experimenter understand what the model is doing and why, to guide the selection of additional model runs and experiments to be carried out, and ultimately to prepare data for use in publications. Besides the intellectual challenges of figuring out what is useful to measure and display, which tend to be somewhat dependent on the specific "content" of each model, there are also a number of technical challenges which often are common across models. For example, for typical agent/individual-based models (ABM or IBM) of Complex Adaptive Systems (CAS), an "experiment" typically consists of running the model many times with the same set of parameter values, just starting with different random number generator (RNG) seeds, to get a distribution of outcomes for that one "case". Then one typically will run many cases, each with different combinations of parameter values, to do sensitivity analysis or just to explore the space of possible parameter values (or initial conditions). Given all that data, it would be nice to have tools to calculate statistics across runs, within runs, for parts of runs, etc., and repeat for a set of cases (e.g., results for different values of a parameter), and create plots of various kinds based on the data and statistics.
About
Read more about the short-term and long-term goals of TFACES here.
Download
Examples
Documentation
TFACES
TFACES.analysis
TFACES.experiment.Experiment
TFACES.experiment.Parameter
TFACES.experiment
TFACES.collection.droneLoader
TFACES.experiment.ParameterSweep
TFACES.experiment.Output
TFACES.collection
TFACES.experiment.Run
CSCS gratefully acknowledges the Google Summer of Code (GSoC) program support for interns to implement this Open Source project,
