多层级常规公交区域协调时刻表编制

发布时间：2018-05-21 07:08

本文选题：多层级常规公交 + 区域协调调度　；参考：《昆明理工大学》2015年硕士论文

【摘要】：在多层级常规公交线网优化背景下,为了使公交运营能更好地满足多样化公交出行需求,依据客流时空分布特征,确定合理的调度形式,建立公交时刻表优化模型。层层深入编制出各层级基于客流需求预测的公交时刻表和适用于多层级公交线网优化的区域协调时刻表。首先,分析了影响多层级常规公交区域协调时刻表编制的主要因素,研究了公交客流的时空分布特征和公交线网结构,确定了多层级常规公交区域协调调度的目标和运营调度形式。其次,分析了公交客流数据的采集和审核方法,分别采用BP神经网络和RBF神经网络算法预测并计算得到公交断面客流需求。设计了层级内基于客流预测的单线公交时刻表优化流程和约束条件,并构建了时刻表方案评价模型。结合实例得出,基于BP神经网络和RBF神经网络预测算法得到的断面客流量,编制得到的时刻表方案,相对于优化前分别节省了2.44%和4.80%的运营总成本。第三,依据多层级常规公交线网优化衔接模式,以各层级公交线路的发车间隔和车辆调度形式作为决策变量,从乘客出行时间成本与公交企业运营收益的角度,考虑乘客舒适性、协同发车间隔和企业运能等方面的约束,建立了多目标优化模型。综合分析各种优化算法的特点后,采用遗传算法、粒子群优化算法和遗传粒子群优化算法求解模型。依据实际问题在MATLAB软件中设计求解模型的算法步骤。最后,通过实例分析得到,在求解本文构建的公交时刻表编制模型过程中,模型目标都能够有效收敛,遗传粒子群优化算法的精度和收敛效率都明显高于遗传算法和粒子群优化算法。求解结果方面：遗传粒子群算法求解得到的多层级常规公交区域协调时刻表方案,相对于基于RBF公交客流需求预测编制的时刻表,分别取三种权重值时的方案总成本分别节省了3.48%、5.47%和8.42%。验证了所建模型和优化算法的可行性和适用性。实际中,需要根据给定的运营效益和乘客出行时间成本权重值选定时刻表优化方案。
[Abstract]:Under the background of multi-level conventional bus network optimization, in order to make the bus operation better meet the needs of diversified public transport travel, according to the characteristics of space-time distribution of passenger flow, the reasonable dispatching form is determined, and the optimization model of bus timetable is established. Layer by layer, the bus timetable based on passenger flow demand prediction and the regional coordination schedule for multi-level bus network optimization are worked out. Firstly, the paper analyzes the main factors that affect the compilation of the regional coordination timetable of multi-level conventional public transport, and studies the space-time distribution characteristics of bus passenger flow and the structure of bus network. The objective and operation form of coordinated regional dispatching of multi-level conventional public transport are determined. Secondly, the methods of collecting and checking the bus passenger flow data are analyzed. BP neural network and RBF neural network algorithm are used to predict and calculate the passenger flow demand of public transport section. The optimization flow and constraint conditions of single line bus timetable based on passenger flow prediction are designed and the evaluation model of schedule scheme is constructed. Combined with an example, it is concluded that the total operating cost is 2.44% and 4.80% lower than that before optimization, respectively, based on BP neural network and RBF neural network prediction algorithm. Thirdly, according to the optimal connection mode of multi-level conventional bus network, taking the departure interval and vehicle dispatching form of each level of bus lines as decision variables, from the point of view of passenger travel time cost and public transport enterprise operating income. Considering the constraints of passenger comfort, cooperative departure interval and enterprise capacity, a multi-objective optimization model is established. After analyzing the characteristics of various optimization algorithms, genetic algorithm, particle swarm optimization algorithm and genetic particle swarm optimization algorithm are used to solve the model. The algorithm of solving the model is designed in MATLAB software according to the practical problem. Finally, through the analysis of an example, it is concluded that the model can converge effectively in the course of solving the model of the bus timetable constructed in this paper. The precision and convergence efficiency of genetic particle swarm optimization are obviously higher than those of genetic algorithm and particle swarm optimization. The solution results are as follows: genetic Particle Swarm Optimization algorithm (GPSO) is used to solve the multi-level bus coordination schedule, which is relative to the schedule based on RBF bus passenger demand prediction. The total cost of the scheme was saved by 3.48% 5.47% and 8.42% respectively. The feasibility and applicability of the proposed model and optimization algorithm are verified. In practice, it is necessary to select the timetable optimization scheme according to the given operation benefit and the weight value of passenger travel time cost.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U491.17

【参考文献】