SELECTION OF OPTIMAL SET OF WEIGHTS IN A LAYERED NETWORK USING GENETIC ALGORITHMS

Genetic algorithms represent a class of highly parallel robust adaptiv e search processes for solving a wide range of optimization and machin e learning problems. The present work is an attempt to demonstrate the ir effectiveness to search a global optimal solution to select a decis ion boundary for a pattern recognition problem using a multilayer perc eptron. The proposed method incorporates a new concept of nonlinear se lection for creating mating pools and a weighted error as a fitness fu nction. Since there is no need for the backpropagation technique, the algorithm is computationally efficient and avoids all the drawbacks of the backpropagation algorithm. Moreover, it does not depend on the se quence of the training data. The performance of the method along with the convergence has been experimentally demonstrated for both linearly separable and nonseparable pattern classes.