Title :
A genetic algorithm for task distribution
Author :
Drabe, Thorsten ; Bressgott, Wolfgang
Author_Institution :
SIBET GmbH, Hannover, Germany
Abstract :
A genetic algorithm is presented to assemble tasks to clusters which are performed by neural network modules. Simulations on letter recognition are compared to those obtained by a monolithic network and by a modular architecture with randomly composed clusters. The proposed method proves superior in terms of final convergence speed, generalization and completeness of solutions
Keywords :
convergence; generalisation (artificial intelligence); genetic algorithms; learning (artificial intelligence); neural nets; optical character recognition; convergence speed; generalization; genetic algorithm; letter recognition; modular architecture; monolithic network; neural network modules; randomly composed clusters; simulations; solution completeness; task distribution; Artificial neural networks; Assembly; Convergence; Evolutionary computation; Genetic algorithms; Hardware; Network topology; Neural networks; Pins; Training data;
Conference_Titel :
Evolutionary Computation, 1997., IEEE International Conference on
Conference_Location :
Indianapolis, IN
Print_ISBN :
0-7803-3949-5
DOI :
10.1109/ICEC.1997.592381
