Parallel Genetic Programming on Graphics Processing Units
Douglas A., S Bernardino Heder, J C Barbosa Helio · InTech eBooks · 2012
In program inference, the evaluation of how well a candidate solution solves a certain task is usually a computationally intensive procedure. Most of the time, the evaluation involves either submitting the program to a simulation process or testing its behavior on many input arguments; both situations may turn out to be very time-consuming. Things get worse when the optimization algorithm needs to evaluate a population of programs for several iterations, which is the case of genetic programming.