Abstract: Besides the ability of stochastic global searching of simple genetic algorithm(SGA), the immune genetic algorithm(IGA) draws into the mechanisms which exist in biological immune system such as immune memory, immune regulation and antibody diversity. The deficiencies of gradient descent method include the slow speed of convergence, the problem of local minima and the great influence of initial parameters on the performance of the network. This paper proposes a new algorithm based on the IGA combined with gradient descent, and applies the new algorithm in the training of RBF network. The experimental results show the algorithm performs well, and it performs better than SGA combined with gradient descent.

Key words: radial basis function (RBF) neural networks, immune genetic algorithm, gradient descent