The rtNEAT package contains source code implementing the real-time NeuroEvolution of Augmenting Topologies method. In addition, it includes code for adaptive neurons. The package includes a public license that allows unrestricted free use of the patent pending rtNEAT technology for noncommercial scientific and research purposes. Parties interested in licensing rtNEAT for commercial use may contact
Rick Friedman at the Office of Technology Commercialization, UT Austin,
software@otc.utexas.edu.
The rtNEAT software is based on the original NEAT source code by Kenneth Stanley. Source code included in this package was originally written by Kenneth Stanley (kstanley@cs.ucf.edu) for the Torque Game Engine used in NERO and later
translated into generic form by Ryan Cornelius (lathentar@mail.utexas.edu). The public rtNEAT release package was then put together by Kenneth Stanley. Please contact kstanley@cs.ucf.edu for questions.
Revision History
- v1.0 (2/24/2006) - original release
- v1.0.1 (2/16/2010) - update release (ikarpov@cs.utexas.edu)
The package was updated with minimal fixes to enable it to compile under GCC 4.x and newer compilers. The package is otherwise identical to v1.0.
- v1.0.2 (9/1/2011) - bug fixes (erkin@cs.utexas.edu):
- Fix typo in NNode::depth() causing max_depth to be calculated incorrectly.
- Fix Makefile optimization flag and a few minor memory issues.
- Fix a performance regression with the Markovian double pole balancing experiment, mainly by increasing weight caps.
- Consider failed runs when printing the average number of evaluations.