NOTICE: THIS PROJECT IS ABANDONED!!! The core feature terms functionality is continued in this project: http://code.google.com/p/fterm/

Agent World provides tools for simulating and visualizing multi-agent systems and is designed for testing machine learning applications (and specially focused on Case Based Reasoning ones). It includes support for representing information using the Feature Term formalism, and provides a series of relational machine learning algorithms that can deal with them. The whole project is created in C++, and uses OpenGL as the visualization library to ensure cross-platformness. Also, the core Feature-Term representation formalism has been reimplemented in Java (due to the implicit memory manager, the delete operation in feature terms can be performed in constant time instead of linear, and thus the whole implementation is much more efficient!).

In addition to the core utilities, this project also includes several test applications that I have been working on for my own research

The main developer is Santiago Ontañón Villar, and the work was initially developed during his Ph.D. Thesis at the IIIA-CSIC research lab in Barcelona (Spain), and has been continued since then, adding functionality and improving the code base.

NOTICE: THIS PROJECT IS ABANDONED!!! The core feature terms functionality is continued in this project: http://code.google.com/p/fterm/