Prediction of performance and problem difficulty in Genetic Programming

Published in Projects
Project details
ID 178323
Duration 2013-2016
Funding ConacytNational Commission on Science and Technology – Mexico
Objective Develop new computational strategies for performance prediction and characterization of problem difficulty for Genetic Programming systems, employing machine learning strategies and evolvability indicators.

Bordeaux Segalen University logo smallInria-corporate smallUniversity of Bordeaux and INRIA – France

Trinity College Dublin Arms smallTrinity College Dublin, Ireland

Summary In this project, we have a taken a machine learning strategy to derive predictors of expected performance for GP systems, as depicted in Figure XX. In essence, the strategy proceeds as follows: