Evolutionary Neural Network Based on Immune Continuous Ant Colony Algorithm

Abstract

Combination of the evolutionary optimization algorithm and traditional neural network, one new kind of neural network which is called evolutionary neural network can be generated. To overcome the demerits of previously proposed evolutionary neural networks, combining the immune continuous ant colony algorithm proposed by author and BP neural network, a new evolutionary neural network is proposed. In this new evolutionary neural network, the architecture and connection weights of BP neural network are all optimized simultaneously by immune continuous ant colony algorithm. At last, to verify this new evolutionary neural network, the typical XOR problem is used. And also, the new evolutionary neural network is compared and analyzed with BP neural network, traditional evolutionary neural network based on genetic algorithm and evolutionary neural network based on evolutionary programming. The computing results show that the precision and efficiency of the new evolutionary neural network are all the best.