A Parallel Genetic/Neural Network Hybrid Learning Algorithm for an MIMD Platform
Sun Yajun; Cao Lei; Yu Juebang(University of Electronic Science and Technology of China, Chengdu 610054)
A new algorithm is presented for training multilayer feedforward neural networks by integrating a genetic algorithm with an adaptive conjugate gradient neural network learning algorithm. This parallel hybrid algorithm hasbeen implemented on an MIMD platform. The performance of the algorithm has been evaluated by applying it to two examples in the domain of image recognition. This parallel hybrid neural network learning algorithm presented in this paper has superior convergente properties.