Features

GROWLab has been designed to facilitate the modeling, simulation, analysis, and validation of complex social processes, with a spe­cial focus on geographic conflict research. Its aim is to bring the development of agent-based simulations to the next level of complexity and realism.

The specific aim of this work is to develop a Java class library that follows the tradition of toolboxes for agent-based modelling, which has turned out to be a successful level of abstraction because of its inherent flexibility. Initially, four core components have been implemented to support:

• The seeding of the model with empirical facts (including geo-referenced data) to calibrate the environments and mechanism to the appropriate level of realism;
• The effective modelling of complex network and hierarchical relationships between model actors and the efficient scheduling of their interactions;
  
• The execution of large number of simulation runs on a grid made of many independent computers to test the sensitivity of the models;
  
• The statistical and visual analysis of the state of the system, as well as the unfolding of the processes over time.