基于云遗传的RBF神经网络的交通流量预测

摘要

以神经网络和混沌时间序列理论为基础，提出了一种基于云遗传的RBF神经网络优化算法。该算法利用云模型云滴的随机性和稳定倾向性的特点，由正态云模型的Y条件云发生器实现交叉操作，由基本云发生器实现变异操作，提高了遗传搜索的效率，精简了网络结构。将该算法应用到Logistic混沌时间序列和实测交通流时间序列进行算法的有效性验证，并与传统的RBF算法和遗传算法优化的RBF算法（GARBF）进行比较。仿真结果表明该算法对混沌时间序列和交通流预测的精度有较大提高，从而证明该算法在交通流时间序列预测领域的可行性和有效性。%Based on neural networks theory and chaotic time series theory, an improved RBF neural networks based on cloud genetic algorithm is proposed. In this algorithm, Y-conditional cloud generator is used as the cross operator and basic cloud generator is used as the mutation operator by utilizing the properties of randomness and stable tendency of normal cloud mode, so improve the efficiency of genetic search and simplify the structure of the network. The efficiency of the proposed prediction method is tested by the simulation of time series of Logistic systems and real traffic flow. The simulation results show that the proposed method in the paper has higher precision compared with the traditional RBF neural network and GARBF neural network, so prove it is feasible and effective in the time series prediction of traffic flow.