商业区公共停车场需求预测模型研究
2020-09-21张军冯悍
张军 冯悍



摘 要: 目前各大城市普遍存在停车难的问题,特别是在商业中心区,人、车流量大,以往停车需求预测方法通常采用停车生成率模型,利用土地用地性质及面积进行预测,文中在停车生成率模型研究的基础上,使用PLS方法对传统停车生成率模型进行改进,模型考虑混合用地之间的相互关系,能够更合理地预测停车需求。以哈尔滨香坊区为例,通过现场调查获得8个综合商业区数据,预测了2019年、2020年的工作日与周末停车需求,并以2个代表性商业区进行验证,实验结果表明,相对原始停車生成率模型,基于PLS改进的停车需求预测模型预测更准确,相对误差更低。
关键词: 商业区; 停车需求预测; 公共停车场; 停车生成率模型; 混合用地; 偏最小二乘
中图分类号: TN911.1?34; U121 文献标识码: A 文章编号: 1004?373X(2020)17?0123?03
Abstract: Parking is a common difficult in metropolises at present. The passenger flow and traffic flow are large, especially in CBDs (central business districts). In the previous parking demand forecast methods, the previous parking generation rate model is usually used to predict the demand according to the nature and area of land use. On the basis of the research on parking generation rate model, the PLS (partial least squares regression) method is used to improve the traditional model. In the model, the relationship among the mixed?use lands are taken into account, so the parking demand can be predicted more reasonably. Xiangfang District in Harbin is taken as an example in the research, data of 8 comprehensive commercial districts are obtained by on?the?spot investigation to predict the parking demand on working days and weekends in 2019 and 2020. Two representative commercial districts are used for verification. The experimental results show that, in comparison with the original parking generation rate model, the PLS improved parking demand forecast model is of more accuracy and lower relative error.
Keywords: commercial zone; parking demand forecast; public parking lot; parking generation rate model; mixed?use land; partial least squares
0 引 言
随着居民生活水平的提高,私家车出行体现了城市化水平,但也对城市交通环境提出更高的要求。在城市中心商业区,土地资源紧张,停车需求大,交通环境紧张,“停车难”的问题尤其严重,为了避免交通拥挤问题的发生,需要合理预测商业区停车需求,对区域停车位进行合理规划与布局。
目前,国内外停车需求预测的主要模型有用地分析模型、交通量?停车需求模型、多元回归分析模型、停车生长率模型等。用地分析模型利用土地用途与员工数来估计城区的停车需求,空间分布可信度高,但不适用于远期预测。交通量?停车需求改进模型认为停车需求是区域吸引该车辆的结果,这种模型适用于用地功能均衡、稳定的短期预测。……
