Adaptive strategies for a semantically driven tree optimizer to control code growth
Bart Wyns, Luc Boullart · 2007
In genetic programming many methods to fight growth exist. But most of these methods require one or multiple parameters to be set. Unfortunately performance strongly depends on a correct setting of each of those parameters. Recently a semantically driven tree optimizer has been developed. In this paper two adaptive strategies to choose a reasonable parameter setting for this growth limiter are presented.