UGM-Markov模型在轨交票务分析中的应用研究

发布时间：2018-11-10 11:15

【摘要】：伴随着经济快速发展,城市化规模不断扩大、交通运输量需求增大。作为绿色环保的城市轨道交通,因其准时、快捷,正受到大、中型城市的青睐。轨交客流预测是城市轨道交通建设的一个重要环节,即可把预测结果作为城市轨交设计规划工作的基础,又可为轨交线路安全运营提供未雨绸缪的决策依据,提高社会和经济效益。本文研究和实践了结合灰色系统理论的UGM(1,1)模型和马尔可夫链适用性建模的一种方式,并应用于轨交客流预测。通过理论分析和实验结果,讨论了相关模型的适用性及不足。本文的主要研究工作概述如下： 1、研究了灰色预测理论中最基本的预测模型GM(1,1),从理论上分析了其适用范围和有偏性。对GM(1,1)进行无偏修正后创建生成UGM(1,1)模型,分析了GM(1,1)和UGM(1,1)模型的不足. 2、在UGM(1,1)模型不足的基础上,结合灰色理论与马尔可夫链构建了Grey-Markov模型,通过实验展示了其优越性及其不足。 3、对Grey-Markov模型进行了无偏修正,提出改进的UGM-Markov模型。通过不断更新建模过程中的原始数列来改变状态矩阵,提高数据状态估算,从而提高模型的预测精度。实验结果表明,Grey-Markov模型较UGM(1,1)模型的预测精度有了很大提高,克服了GM(1,1)和UGM(1,1)模型对波动性大的数列预测精度不高的缺陷。改进的UGM-Markov模型通过更新原始数列来提高估算精度,其预测效果最佳。
[Abstract]:With the rapid development of economy, the scale of urbanization is expanding and the demand of traffic volume is increasing. As a green urban rail transit, because of its punctual, fast, is favored by large, medium-sized cities. The prediction of rail transit passenger flow is an important link in the construction of urban rail transit. It can be regarded as the basis of the design and planning of urban rail transit, and it can also provide a decision basis for the safe operation of rail transit line in advance. Improving social and economic benefits. This paper studies and practices a method of UGM (1K1) model combined with grey system theory and the applicability of Markov chain, and applies it to the prediction of rail passenger flow. Through theoretical analysis and experimental results, the applicability and deficiency of the related models are discussed. The main research work of this paper is summarized as follows: 1. The most basic prediction model GM (1K1) in grey prediction theory is studied, and its applicability and bias are analyzed theoretically. After unbiased modification of GM (1K1), the UGM (1K1) model was created and the shortcomings of GM (1K1) and UGM (1K1) models were analyzed. 2. On the basis of the insufficiency of UGM (1K1) model, the Grey-Markov model is constructed by combining grey theory with Markov chain. The advantages and disadvantages of the model are demonstrated by experiments. 3. The Grey-Markov model is modified unbiased, and an improved UGM-Markov model is proposed. By constantly updating the original sequence in the modeling process, the state matrix is changed and the data state estimation is improved, thus the prediction accuracy of the model is improved. The experimental results show that the prediction accuracy of the Grey-Markov model is much higher than that of the UGM (1K1) model, which overcomes the defects of the GM (1t1) and UGM (1K1) models in the prediction accuracy of the series with large volatility. The improved UGM-Markov model improves the estimation accuracy by updating the original sequence, and its prediction effect is the best.
【学位授予单位】：东华大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：O211.62;U293.13;U239.5

【相似文献】