北京市摩拜共享单车源汇时空特征分析及空间调度

共享单车是解决城市出行最后一公里的重要工具。然而,共享单车在使用过程中出现了供需时空失衡的现象,导致部分区域无车可用,而另一些区域却车满为患。这一现象不仅给用户带来不便,同时也降低了单车周转效率。解决供需失衡问题,关键在于探究共享单车供需失衡的时空分布特征。本文定义了共享单车的“源”、“汇”点,采用2017年5月10日至2017年5月16日北京市摩拜单车数据以及土地利用分类等多源数据,分析了共享单车工作日与周末,以及早、晚高峰期间强源、强汇点的分布特征,并结合土地类型信息分析了不同用地类型的单车使用模式,从而进一步提出了共享单车的空间调度模型。结果显示,在考虑不同土地利用类型的影响下,摩拜共享单车强源汇点分布模式具有明显的时空异质性：① 工作日单车使用量明显高于周末,且不同土地利用类型的源汇分布显著不同,如居住用地、商业金融用地等地净流入、流出密度更大,绿地等地区则相对较小;② 对比早晚高峰期间,同一地区的单车使用源汇模式极可能相反,如教育科研用地、商业金融用地等带有办公性质的地区会有“早汇晚源”的特征,而居住用地则是“早源晚汇”;③ 同一类土地利用在工作日与周末的早晚高峰期间,单车使用的源汇特征亦存在差异,如办公性质的地区在周末时源汇比例会明显产下降。基于上述结果,本文提出了一种局部优化的调度模型,并通过实际数据进行了检验。该模型在一定程度上可以解决车辆空间分配不均衡的问题,提高城市共享资源使用率,增加人们出行的方便程度。

Spatial-temporal Characteristics of Source and Sink Points of Mobikes in Beijing and Its Scheduling Strategy

Sharing bicycles is an important tool for solving the "last mile" travel problem in a city. However, the imbalance between supply and demand of sharing bicycles occurred frequently, which leads to significant spare of bicycles and obvious inconvenience for users. The key to solve this problem is to understand the spatial and temporal distribution characteristics of sharing bicycles’ supply and demand. In this case, this study defines the intensive "source" and intensive "sink" points of sharing bicycles, and uses Mobikes cycling data and land use type data to analyze the distribution characteristics of the intensive source and intensive sink points during the workdays and weekends, as well as the morning and evening peaks. Combined with the land type information, we further analyze the characteristics of sharing bicycles' usage under different land use types and propose a local scheduling model based on spatial neighborhood. Results show that under different land use coverage, the distribution of intensive source and intensive sink points of sharing bicycles show significant spatial and temporal heterogeneous patterns: (1) The usage of sharing bicycles during work days is obviously higher than weekends, and the distribution of the intensive source and intensive sink is obviously different among different land use types. For example, the net inflow and outflow density in areas such as residential, commercial, financial lands are greater than those of green areas; (2) In the morning and evening peak periods, the source and sink attributes of the same area are always showing a converse pattern. For instance, areas with offices such as education and scientific research lands and commercial and financial lands will have the characteristics of "morning sinks and evening sources", while the residential lands will be "morning sources and evening sinks"; (3) During the morning and evening peak periods of weekdays and weekends, there are also differences in characteristics of the sources and sinks of the same type of land use. For instance, the proportion of the intensive sources and intensive sinks will obviously decline in the office area on weekends. Based on the results above, this paper proposes a local optimized scheduling model. It is found that the model can effectively reduce the imbalance of bicycles' space allocation and increase the resource's utilization and the convenience of people's travelling.