基于萤火虫算法优化BP神经网络的公交行程时间预测

发布时间：2018-05-22 08:00

本文选题：智能交通系统 + 公共交通　；参考：《广西师范大学学报(自然科学版)》2017年01期

【摘要】：通过对公交车辆行驶特性以及行驶环境的实际分析,本文确立了影响公交车辆行程时间的几个关键因素:天气状况、时间段(高峰/平峰)、交通流量以及路段长度。结合BP神经网络具有的非线性映射能力、自学习和自适应能力、泛化能力强和萤火虫算法的参数少、操作简单、稳定性好等优点,本文提出一种利用萤火虫算法对BP神经网络进行优化的算法,以减少神经网络的训练时间和提高其预测稳定性。然后利用GPS数据和实时采样数据对神经网络进行训练和建模,实现公交车辆行程时间的准确预测。将该优化算法与传统的BP神经网络算法和卡尔曼滤波算法用MATLAB分别仿真测试,对比结果显示,该算法对公交行程时间的预测具有更高的准确性且结果更加稳定。
[Abstract]:Based on the actual analysis of the driving characteristics and driving environment of public transport vehicles, several key factors affecting the travel time of public transport vehicles are established in this paper: weather condition, time period (peak / peak, traffic flow and length of section). The BP neural network has the advantages of nonlinear mapping ability, self-learning and adaptive ability, strong generalization ability, less parameters of the firefly algorithm, simple operation, good stability and so on. In this paper, an algorithm for optimizing BP neural network using firefly algorithm is proposed to reduce the training time and improve the prediction stability of neural network. Then GPS data and real-time sampling data are used to train and model the neural network to realize the accurate prediction of bus travel time. The optimization algorithm is compared with the traditional BP neural network algorithm and the Kalman filter algorithm respectively by MATLAB simulation. The comparison results show that the algorithm has higher accuracy and more stable results in the prediction of bus travel time.
【作者单位】：华南理工大学土木与交通学院;
【基金】：国家自然科学基金(51308227) 广东省交通运输厅科技项目(科技-2015-02-076)
【分类号】：U491.17;TP18

【相似文献】