基于公交大数据的高平峰预测与调度优化设计

Peak and Ordinary Time Forecast and Dispatch Optimization Design Based on the Big Data of Public Transport

从传统的“交通工程”到今天的“智慧城市”,以某公交线路为例,通过测算公交出行大数据,实现对该线路的调度优化,旨在为今后的交通规划设计提供科学参考。首先,采用k-means聚类来定义合理的公交高平峰;其次,采用决策树将高峰事件继续细分为3个子事件,运用风险生存函数预测公交高平峰时间段并进行检验;最后就深圳17路公交车路线数据进行实证。根据预测结果,以公交公司运营的总车辆数最小为目标,在公交公司的利益反映为车辆满载率和运营成本,以及乘客的利益反映为乘客总等待时间的约束条件下,构建整数规划模型测算其调度方案。

From the traditional“traffic engineering”to today’s“smart city”,taking a certain bus route as an example,the dispatch optimization of this bus line is realized through analyzing the corresponding big data,aiming to provide a scientific reference for future transportation planning and design.First,k-means clustering is used to define a reasonable peak and ordinary time of public transportation.Secondly,a decision tree is used to further subdivide the peak event into three sub-events,and the risk survival function is used to predict and test the time period of the high peak of public transportation.Finally,an empirical study is conducted based on the route data of Bus 17 in Shenzhen.According to the forecast results an integer programming model is constructed to calculate the dispatching plan with the goal to minimize the total number of vehicles operated by public transport companies under the constraints of the vehicle full load rate and ope-rating cost in terms of the interests of public transport companies and the total waiting time in terms of the interests of passengers.