摘要
针对高速公路交通量与其经济影响因素之间的复杂非线性关系,将最小二乘支持向量机(least squares support vector machines,LSSVM)与自适应动态粒子群优化(adaptive dynamic particle swarm optimization,ADPSO)算法相结合,提出一种ADPSO算法优化LSSVM的高速公路交通量新型预测方法.将建模简单、精度高的LSSVM作为预测模型,通过寻优能力优异的ADPSO算法选择LSSVM最优参数.以某市高速公路交通量为例验证模型的有效性.结果表明,所提方法的预测性能较好,适合于高速公路交通量的短期预测.
There is a complex nonlinear relationship between highway traffic flow and its influencing factors. Combing least squares support vector machines （LSSVM） with adaptive dynamic particle swarm optimization （ADPSO） algorithm, this paper proposed a new highway traffic flow forecasting method based on LSSVM optimized by ADPSO algorithm. Highway traffic flow was forecasted by LSSVM with the advantages of easy modeling and high precision. And the optimal parameters of LSSVM were selected based on the good optimization ability of ADPSO algorithm. An example analysis on the highway traffic flow in a city was performed to test the effectiveness of LSSVM-ADPSO model. The results indicate that the proposed method has better highway traffic flow forecasting performance and is suitable for short-term highway traffic flow forecasting.
出处
《河北大学学报（自然科学版）》
CAS
北大核心
2017年第3期302-308,共7页
Journal of Hebei University(Natural Science Edition)
基金
国家自然科学基金青年项目(61503261)
河北省交通运输厅科技计划项目(Y-2010024)
北华航天工业学院科研基金项目(KY-2015-09)
河北省软科学研究计划项目(15456106D)
河北省高等学校青年拔尖人才计划项目(BJ2014097)
河北省社会科学发展重点课题(2015020206)
国家留学基金委(CSC)公派留学地方合作项目(201608130165)
河北省高校人文社会科学重点研究基地石家庄铁道大学工程建设管理研究中心资助项目
河北省软科学工程建设管理研究基地资助项目
河北省重点学科管理科学与工程资助项目
关键词
高速公路
交通量预测
自适应动态粒子群优化算法
最小二乘支持向量机
highway
traffic flow forecasting
rithm
least squares support vector machines adaptive dynamic particle swarm optimization algo
