基于灰色神经网络的航线客流量预测

准确预测航线客流量对于航空公司制定航线销售政策有着重要的作用。现有研究中鲜见考虑民航旅客出行的随机性、客流量表现出的非线性特征以及对航线客流量影响因素的分析。针对以上问题,提出一种基于灰色神经网络的航线客流量预测模型。该模型运用灰色理论弱化数据序列的随机性,再结合非线性处理能力较强的BP神经网络,构建基于灰色神经网络的航线客流量预测模型。同时验证了平均折扣率对航线客流量的影响。实验结果表明,相比于灰色GM(1,2)模型、BP神经网络模型,灰色神经网络模型具有更高的航线客流量预测精度和更强的稳定性。     

基于交变粒子群BP网络的电力系统短期负荷预测

短期负荷预测是电力系统正常运行的关键环节,合理的发电计划依靠准确的负荷预测,因此提出交变粒子群算法来优化BP网络模型以预测电力短期负荷。针对 依靠先前的经验 来确定BP神经网络的权值缺少理论依据的问题,采用交变粒子算法优化BP神经网络权值,以减少通过神经网络预测模型求解电力短期负荷预测带来的误差。实验证明,经过优化的BP神经网络预测模型比传统的BP神经网络预测模型的误差更小,更加接近实际电力负荷。     

Inefficiency of Logit-Based Stochastic User Equilibrium in a Traffic Network Under ATIS

This paper makes a contribution to the literature by bounding the travel time inefficiency of the logit-based stochastic user equilibrium (SUE) under Advanced Traveler Information Systems (ATIS). All drivers are divided into two groups, one equipped with ATIS and another without, and both of which follow the logit-based SUE principle in making route choices. The equipped drivers have less degree of travel time variability than the unequipped ones. The inefficiency of the two-user class SUE is defined in two different ways, i.e., in comparison with the SO in terms of total actual system travel time, or in comparison with the corresponding SSO in terms of total perceptive system travel time of all users. The effects of various parameters on the bounds are further investigated. It is found that the inefficiency bound against the SSO is only dependent upon the degree of link congestion and independent of the network topology. In contrast, besides the effect of the degree of link congestion, the increasing of total demand and network complexity will also make the inefficiency bound against the SO go up, while the promotion of ATIS market penetration and perception benefit will reduce the bound.     

高速公路动态交通流的BP神经网络建模

通过对高速公路宏观动态交通流模型的分析,针对高速公路交通系统的非线性时变特点,应用BP神经网络建立了高速公路宏观动态交通流模型。并利用一段高速公路的交通流数据对BP神经网络进行训练,得到网络参数。最后,为了验证BP网络模型的有效性,在MATLAB环境中对模型进行了仿真,并将仿真结果与原始模型的结果进行了比较。结果表明,该方法能较准确地描述高速公路交通流的真实行为,并且能够适应交通状况的变化。     

Comparing paradigms for strategy learning of route choice with traffic information under uncertainty

This paper aims to model the traveller's day-to-day route choice in the case of an Advanced Traveller Information System (ATIS) through two learning paradigms: reinforcement-based and belief-based. The reinforcement learning approach is adopted in both a basic version and an extended one. Similarly, the belief-learning approach is adopted in both a Joint Strategy Fictitious Play version and in a Bayesian-learning one. All the models are compared and validated based on data collected by means of a stated preference experiment. The models explicitly account for the reliability of the information system, as this interacts with the inherent dispersion of network travel times and determines the overall level of uncertainty affecting the travellers’ adaptive learning behaviour. The experiment is then designed to simulate different levels of reliability for the ATIS. Results show that for intermediate and high levels of information accuracy, joint strategy fictitious play best predicts the respondents’ route choice behaviour under information provision, suggesting that a best-reply strategy is used by travellers for their route choices. In low information accuracy, the result suggests the payoff variability moves the choice behaviour toward randomness. The proposed approach provides useful tools to model travellers’ adaptive route choice behaviour and contributes to the support of effective ATIS design.     

基于BP神经网络的预测建模系统的研究与实现

神经网络具有良好的记忆、归纳和学习能力，对难以用数学方法建立精确模型的信息、工艺等能够进行有效地预测建模。该文通过对BP神经网络的分析和研究，针对传统BP算法的不足，采用Levenberg—Marquardt(LM)优化算法的建立一个基于BP神经网络预测建模系统。在介绍了系统的主要功能之后，给出了用MATLAB软件实现该系统主要模块的具体程序。最后采用该系统对一个制造过程中刀具磨损量的进行了预测建模，实验仿真结果表明：系统具有良好的预测效果，刀具实际磨损量与预测磨损量的误差基本上在10％以下。     

带反馈输入BP神经网络的应用研究

为了有效解决具有非线性特征的水文预报精准度的问题,通过对反向传播BP神经网络的学习和研究,分析了变量间的相互信息,提出了系统间相关信息熵的概念,并建立了适合水文预测的自迭代反向传播神经网络模型.该模型通过对迭代因子的及时修正,在反向传播中不断调整网络的权值和阈值,从而在很大程度上改善了传统BP算法所带来的不足,提高了预测的精度.实际的应用研究表明,自迭代反向传播模型的预测效果优于传统预测模型.     

Application of fuzzy logic and neural networks for dynamic travel time estimation

The effectiveness of loop detectors as a data source for advanced traveler information systems has been researched recently [V. P. Sisiopiku (1993 Travel Time Estimation from Loop Detector Data for Advanced Traveler Information System Applications , Ph.D. Thesis, University of Illinois at Chicago]. In urban traffic control schemes loop detectors provide on-line information on traffic conditions consisting of volume counts and occupancy levels. The need to convert loop detector data into travel times is recognized mostly in data fusion applications [P. Nelson and P. Palacharla (1993) A neural network model for data fusion in ADVANCE, Pacific Rim Transportation Technology Conference Proceedings , Vol. I, pp. 237–243, Seattle, WA, 1993]. Literature review indicates limited knowledge on the actual relationship between travel times and loop detector data under interrupted traffic conditions [V. P. Sisiopiku and N. M. Rouphail (1994) Towards the Use of Detector Output for Arterial Link Travel Time Estimation: a Literature Review. Transportation Research Record Series , Washington, DC]. Currently available statistical regression models cannot capture the dynamics of traffic conditions under signalized control and suffer from limited calibration and empirical validation. This paper presents a fuzzy reasoning model to convert loop detector data into link travel times obtained from empirical studies. This model incorporates flexible reasoning and captures non-linear relationship between link specific detector data and travel times.     

面向高速公路通行时间分布预测的时空混合密度神经网络

准确的通行时间分布预测可以全面地反映高速公路路网中各个路段在未来的通行状况,辅助实现高速公路中的路径规划,事故事件预警等精细化管理目标.为此,本文提出一种面向高速公路通行时间分布预测的时空混合密度神经网络.具体地,本文利用自适应图卷积通过数据驱动的方式提取路网中的空间特征,有效解决了基于预定义图难以捕获路网信息中完整空间相关性的问题.在时间维度上,不同时间的路网信息存在显著的相关性,因此,本文基于注意力机制自适应建模路网信息的时间相关性,并通过卷积层进一步聚合相邻时间步之间的信息.最后,基于自适应时空相关性建模得到的路段嵌入表示,通过混合密度网络建模通行时间的分布,以实现高速公路中各个路段的通行时间分布预测.     

基于改进遗传算法优化BP神经网络的销售预测模型

随着经济的快速发展,众多企业步入科学化管理的时代.销售预测是企业经营活动中必不可少的一个环节,预测的准确性直接关系到销售经营的成败.因此提出基于传统BP神经网络与时间序列预测模型为一体的改良BP神经网络预测模型,通过该模型的预测,可以更可靠地预测企业在未来单位时间内的销售额.改良神经网络参考了同步时间序列的预测做出了自我校准,并利用遗传算法达到通过校准得到自我优化的目的,简化网络结构,提高预测的准确度.