"One of the problems faced in applying neural network to some real
world application is related to difticulties in finding an optimum set of weights
and thresholds during the training phase. A general most method in tinding
these solutions for these problems is backpropagation.
A different method to tind the solutions of the same problems is
Genetic Algorithms. Genetic algorithm is relatively new search algorithm that
has not been fully explored in this area. ln this thesis, genetic algorithms are
applied to train neural networks and to evolve an optimum set of weights and
thresholds. Process begin with encode neural networks parameters to binary
chromosomes, and evaluate. The Spinning wheel selections are using to
produce offspring with high titness_ then recombinate with crossover and
mutation as genetic operator.
The proiect carried out investigates whether genetic atgonthms can be
applied to neural networks to solve pattem classitication and function
approximation problems. This thesis describes tl1e simulation works that
have been perfomwed. It describes the design ofa genetic algorithm and the
results obtained. ln pattem classilication problem that use feedforward
network show, that genetic algorithm is superior to backpropagation training
rule in error and speed calculation. ln function approximation, the result
shows that genetic algorithm approach is very much slower than the
backpropagation method. Results' show that even for relatively simple
network, genetic algorithm requires a much longer time to Uain neural
networks-"
