Hybrid Intelligent Algorithm of RBF Neural Network with Variable Length Structure

作者：She Yuanguo^*

来源：Information-An International Interdisciplinary Journal, 2012, 15(6): 2571-2577.

• Gauss-Newton descend method
• RBF neural network
• Hybrid intelligent algorithm
• Variable length structure

Based on hybrid encoding, a hybrid intelligent algorithm for the optimization of RBF neural network is constructed in this paper. This method designs the Chromosome of genetic algorithm with variable length structure, takes genetic algorithm which combined with Gauss-Newton descend search, immune characteristics and chaos idea to optimize the RBF network structure and hidden layer parameters, takes least squire method (LSM) based on singular value decomposition to optimize the linear weights directly. Simulation results demonstrate it is effective and superior to some other methods.

• 出版日期2012-6
• 单位中国人民武装警察部队工程大学