公交运力资源优化配置的分析与研究

【作者】 周青；

【导师】 吴晟；

【作者基本信息】 昆明理工大学 ， 计算机系统结构， 2012， 硕士

【摘要】 面对城市日益增长的客流需求,最大限度地优化公交运力资源配置势在必行。改善公交运力配置,不权仅是提高公交运营效率的需要,更是提高公交服务水平,实现公交运力资源均衡配置和创建环境友好型社会的内在需要。要解决公交问题,仅仅依靠增加公交车的数量,或者调整公交线路并不能从根本上解决问题,相反,一味地增加公交车的数量,不权会造成公交资源的浪费,更会加大城市交通的压力。因此,只有充分地利用现有的公交资源,在满足公交企业和乘客多方利益的前提下,对相关资源进行优化配置,才是解决问题的关键所在本文首先在综合分析公交运力配置的影响囚素的基础上,将公交运力优化配置问题定性为多目标优化问题,并进一步介绍了基于基尼系数的多目标优化模型,在分析原模型的基础上,引进了基尼系数这一经济学指标,通过量化表示将资源配置均衡度作为新的目标函数加入原模型,得到了改进的新模型。并通过实例化数据验证,通过前后的调整,各站点的客流量分担率(基尼系数)由原来的0.29改进为0.21,从客流量分布均衡性的角度说明了新模型比原模型具有较好的现实拟合度。本文明确了客流量预测在整个公交运力配置流程中的关键作用。客流量的准确预测是公交资源优化配置的依据和前提。本文给出了基于Elman神经网络的公交客流量预测流程。最后,通过数据搜集,利用北京市某条线路近8年的客流量数据进行了实例预测。最后通过与当前较流行的BP静态神经网络比较,发现Elman神经网络整体上比BP神经网络具有较高的预测精度。更多还原

【Abstract】 With the continuous growth of the city traffic. optimizing bus resources as possible as can is imperative. To improve the public transport capacity configuration, is not only able to help improve public transport operation efficiency, but also improve the bus service level, realize the balance of whole bus transport resource and create environment-friendly society.To solve the bus problems, only relying on increasing the number of public buses or adjusting the bus lines, can’t resolve the problem ultimately, instead, increasing the number of public buses blindly, not only can cause public transport waste of resources, but also will increase the urban traffic pressure. Therefore, how to make full use of the existing public transport resources, and how to meet the interest of bus passengers and State Transit,under the premise of related resources to optimize the allocation, is the key to the problem.First of all, the factors that relevant factor of public transport capacity are comprehensively analysed so we characterize optimal allocation of public transport capacity as multi-objective optimization problem in this paper. Moreover. multi-objective optimization model is introduced that bases on game theory. By analysis of the original model, we introduce the Gini coefficient in the economic indicators. After quantization, the degree of resource allocation equilibrium is viewed as a new objective function by putting in the original model. A new and improved model is obtained finally. Applying to instance data. by adjustment, the share rate of the passenger volume (Gini coefficient) improved from0.29to0.21.we can conclude that the new model is better than the original model in line with reality from the balance of traffic distribution。This paper made clear that bus passenger forecast plays the key role in the entire capacity configuration flow. Accurate prediction of the passenger is the basis and premise of bus resources optimization allocation. In this paper.the bus passenger forecasting model based on the liman neural network is given. Finally.the forecast is carried on using the nearly eight years data of Beijing passengers. Finally by comparing with BP static neural network which is more popular currently, we discover that Elman neural network is of high precision than the BP neural network.更多还原