Mathematical Methods for Assisting Agent-Based Computation
The MMAAC project is an attempt to overcome computational efficiency issues common in complex social simulation. Together with Ioannis G. Kevrekidis's group at Princeton, CASOS is working to integrate Kevrekidis's "Equation-Free Framework" with our multi-agent simulation tool Construct. Originally developed as a tool used in complex/multiscale simulations of chemical and biological systems, Equation-free Modeling allows for expedited computation and analysis of these systems. Specifically it is most appropriate for use in systems in which accurate models only exists for local (micro-level) components but system (macro-level) behavior is the phenomena of interest for the researcher. This perspective easily extends to social systems, while we have strong sociological and cognitive models for how people think and act in social systems we lack accurate and robust system level models. The augmentation of Construct with the Equation-Free Framework could allow for more computationally efficient social simulation with a minimal loss in predictive capabilities.