摘要
针对短时交通流预测模型存在稳定性不好、预测精度不高等问题,提出一种利用改进的花授粉算法(SFPA)优化支持向量机的短时交通流预测模型。将短时交通流时间序列进行相空间重构,利用模拟退火算法对FPA进行改进,形成SFPA优化算法,并应用到支持向量机参数优化过程中,构建一种SFPA-SVM短时交通流预测模型。通过实例对该模型进行性能测试,测试结果表明,SFPA-SVM模型提高了预测精度,对短时交通流预测是有效可行的。
There are some problems,such as low prediction,in short-term traffic flow forecast model.To solve these problems,the short-term traffic flow forecast model of support vector machine（SVM）optimized by improved flower pollination algorithm（FPA）was proposed.The time series of short-term traffic flow were transformed,the simulated annealing（SA）was used to improve flower pollination algorithm,and this optimization algorithm（SFPA）was used to optimize the parameters of support vector machine and a SFPA-SVM short-term traffic flow prediction model was built.The performance of this model was tested by the simulation experiments.The simulation results show that the SFPA-SVM model improves the prediction accuracy,and it is suitable and effective for forecasting short-term traffic flow.
出处
《计算机工程与设计》
北大核心
2016年第10期2717-2721,共5页
Computer Engineering and Design
基金
江苏省普通高校研究生科研创新计划基金项目(SJLX_0334)
江苏省科技厅软科学基金项目(BR2012043)
关键词
短时交通流
相空间重构
模拟退火
花朵授粉
支持向量机
short-term traffic flow
phase space reconstruction
simulated annealing
flower pollination algorithm
support vector machine
