GeneticImprovement

Genetic Improvement

Genetic Improvement (GI) is an application of GP (and Linear GP), aimed at automatically patch or tweak existing code for different purposes: correct bugs, reduce memory usage, improve performance with regards to a specific case study, ... In this example, MicroGP is used to create a language-independent GI system. The resulting framework is then tested on a few case studies. This application represents Francesco Marino's Master Thesis Project.

For more information on Genetic Improvement, see:

1. http://geneticimprovementofsoftware.com/ (Annual workshop on GI)
2. http://www0.cs.ucl.ac.uk/staff/ucacbbl/gismo/ (GISMO, software tool for GI on multiple objectives)

Source for the project

Constraints, evaluator, configuration for a simple case study (targeting the same code in C++, Python and Java). Simply unpack all the files in a folder, and run MicroGP in that directory.

Publications

MARINO F., SQUILLERO G., TONDA A., A General-Purpose Framework for Genetic Improvement, In: Proceedings of PPSN 2016, 14th International Conference on Parallel Problem Solving from Nature