Title :
Using fitness distributions to improve the evolution of learning structures
Author :
Igel, Christian ; Kreutz, Martin
Author_Institution :
Inst. fur Neuroinf., Ruhr-Univ., Bochum, Germany
Abstract :
The absolute benefit, a measure of improvement in the fitness space, is derived from the viewpoint of fitness distribution and fitness trajectory analysis. It is used for online operator adaptation, where the optimization of density estimation models serves as an example. A new information theory based measure is proposed to judge the accuracy of the evolved models. Further, the absolute benefit is applied to offline analysis of new gradient based operators used for coefficient adaptation in genetic programming. An efficient method to calculate the gradient information is presented
Keywords :
genetic algorithms; information theory; learning (artificial intelligence); probability; absolute benefit; coefficient adaptation; density estimation models; fitness distributions; fitness space; fitness trajectory analysis; genetic programming; gradient based operators; gradient information; information theory based measure; learning structure evolution; offline analysis; online operator adaptation; Algorithm design and analysis; Density measurement; Distributed computing; Evolutionary computation; Gain measurement; Genetic communication; Genetic programming; Information theory;
Conference_Titel :
Evolutionary Computation, 1999. CEC 99. Proceedings of the 1999 Congress on
Conference_Location :
Washington, DC
Print_ISBN :
0-7803-5536-9
DOI :
10.1109/CEC.1999.785505
