In this paper the efficiency of adaptive mutations in the genetic algorithms used for training radial basis network and optimisation of its structure is considered. We investigate the following types of adaptive mutations: decreasing Cauchy mutation, adaptive Gauss and Laplace mutations. A comparative analysis using simulation is conducted.
neural network, genetic algorithm, adaptive mutation, basis function, selection, fitness function, chromosome, gene
"Obuchenye radyalno-bazysnыkh setei s pomoshchiu henetycheskykh alhorytmov s adaptyvnoi mutatsyei" ,
Information Processing Systems,