基于多级目标规划的高速公路二义性路径识别

Multi-objective programming based ambiguous routes identification of expressway network

高速公路二义性路径的准确识别是保证联网收费公平分配的关键技术。从路网流量的客观约束条件和用户主观行为描述两个角度出发,提出基于车辆检测器客观监测数据的多级目标规划识别方法,并融合已有的描述用户主观行为的概率识别方法,建立了多级目标规划模型,同时发挥了两种方法各自的优势,提高了识别准确度。采用SUMO交通仿真软件对其进行模拟评价,结果表明:在至少50%车检器能够正常工作的情况下,流量约束识别方法的识别准确度较已有概率方法提高约13%,多级目标规划模型的识别准确度较流量约束识别方法再提高约10%.

Ambiguous routes identification in expressway plays an important role in fair distribution of network toll collection. In this paper, a sensor data based multi-objective programming method was proposed based on traffic flow objective constraints and users’ subjective choice. In addition, by integrating existing probability methods which describe users’ choice, multi-objective programming model is proposed to take advantages of both kind of methods and improves the identification accuracy. By using SUMO traffic simulator for evaluation, experiments show that with at lest 50% of the sensors working normally, the flow constraint method’s accuracy is about 13% higher than existing methods, and accuracy of multi-objective programming model is improved than that of flow constraint method at about 10%.