基于多源数据的公交专用道效能评价方法与影响模型

施画公交专用道是落实公交优先策略、改善公交服务质量,提升交通系统效率的重要策略.近年来,如何提高公交专用道设置的科学性、提升服务效能受到广泛关注.为科学定量评价公交专用道设置的正、负效能,建立基于多维运行影响的公交专用道效能评价指标体系,提出基于主观层次分析法和客观离差最大化法相结合的组合评价方法,实现对公交专用道效能...

Evaluation Method and Influence Model of Bus Lane Performance Based on Multi-source Data

The bus lane is an important factor responsible for implementing bus priority strategy, and improving the quality of bus services and efficiency of traffic systems. In recent years, improvements in the scientific setting and service efficiency of bus lanes have been widely studied. To evaluate the positive and negative performance of bus lanes quantitatively and scientifically, and to explore the influencing factors of the performance of bus lanes, this study established an indicator system of bus lanes performance evaluation based on the multi-dimensional operation influence, and proposed a combined evaluation method based on the combination of the subjective analytic hierarchy process and objective deviation maximization to realize quantitative evaluation of bus lanes. Consequently, the influencing factors of bus lane performance were selected and determined using grey correlation analysis; the influencing model based on the random forest regression model was established; and the influence degree of each factor on bus lane performance was measured by the evaluating the importance of variables. Based on the multi-source traffic operation data, this study used the Beijing-Tibet Expressway, Beijing-Hong Kong-Macao Expressway, and Third Ring Roads as examples to verify the model. The results show that the proposed model can accurately describe the performance of bus lanes under different time and space conditions; The results demonstrated that the type of bus lane has the greatest impact on its performance, followed by the road saturation, bus vehicle volume, break points density, section speed, and bus passenger flow. The model comparison shows that the evaluation accuracy of the proposed model based on random forest regression is better than that of the logistic regression model and back propagation neural network model. Relevant research can provide important method support for the fine design and scientific setting of bus lanes.