Multi-city traffic flow forecasting via multi-task learning

Applied Intelligence - Traffic flow forecasting or prediction plays an important role in the traffic control and management of a city. Existing works mostly train a model using the traffic flow...     

基于核学习方法的短时交通流量预测

基于核学习的强大非线性映射性能,针对短时交通流量预测,提出一类基于核学习方法的预测模型。核递推最小二乘(KRLS)基于近似线性依赖(approximate linear dependence,ALD) 技术可降低计算复杂度及存储量,是一种在线核学习方法,适用于较大规模数据集的学习;核偏最小二乘(KPLS)方法将输入变量投影在潜在变量上,利用输入与输出变量之间的协方差信息提取潜在特征;核极限学习机(KELM)方法用核函数表示未知的隐含层非线性特征映射,通过正则化最小二乘算法计算网络的输出权值,能以极快的学习速度获得良好的推广性。为验证所提方法的有效性,将KELM、KPLS、ALD-KRLS用于不同实测交通流数据中,在同等条件下,与现有方法进行比较。实验结果表明,不同核学习方法的预测精度和训练速度均有提高,体现了核学习方法在短时交通流量预测中的应用潜力。     

基于浮动车数据非参数回归短时交通速度预测

非参数回归算法是近年来提出的一种较新型的短时交通流预测算法,为了提高预测精度,提出了基于误差反馈的预测方法.加入误差反馈机制,针对状态向量中的权值进行实时的反馈修改,得到了较满意的结果.采用成都市浮动车系统中的出租车浮动车数据对红星路二段的速度进行了预测,预测结果表明,该算法的预测精度优于无反馈的非参数回归和BP神经网络.     

融合时空信息的短时交通流预测

为了准确描述交通流的时空演化过程并提高交通流短时预测的精度,融合时空交通流信息,即时间维度的交通流量信息和空间维度的路网耦合信息,构造基于GM（1,N）-Markov 链的组合预测模型。将预测路段与关联路段看作是一个灰色系统并对其进行灰关联分析,通过对灰关联度最低阈值的设定,实现了空间信息的深度挖掘和对无效信息的过滤清洗;利用多维GM（1,N）模型对预测点与强关联点作全局、系统的分析预测,并针对GM（1,N）对随机性较大的数列可能出现预测失效的问题,引入马尔科夫链对模型进行修正;利用VISSIM对模型进行仿真验证,分别以2 min、5 min、10 min为时间间隔进行仿真模拟,预测平均相对误差分别为9.30%、5.95%、3.20%,模型精度均为优,证实模型是有效的。     

基于CNN-XGBoost的短时交通流预测

为准确预测短时交通流,缓解交通拥堵提高交通运行效率,提出一种基于CNN-XGBoost的短时交通流预测方法。结合短时交通流数据的时间相关性和空间相关性,将本路段和邻近路段的历史数据一同作为输入进行预测。利用卷积神经网络(convolutional neural networks,CNN)实现特征提取以减少数据冗余性,提出一种参数经果蝇算法优化的XGBoost模型用于交通流量预测。实例验证结果表明,CNN可对时间和空间结合下的交通流数据进行有效特征提取;相比SVR、LSTM等模型,改进的XGBoost模型下的交通流量预测误差明显减小。     

处理缺失数据的短时交通流预测模型

针对交通检测中数据的缺失问题,提出了一种新的交通流综合短时预测模型,这种模型可以对交通检测中的缺失数据进行重建,并在此基础上运用改进的卡尔曼平滑算法进行短时交通流预测。该模型克服了传统的预测方法无法对检测数据的缺失进行处理的缺点,能在数据缺失时进行有效的交通流预测。通过深圳市的实际流量数据的验证,并比对传统方法,证实该方法具有较好的预测性能,模型预测精度可以保持在88%以上,具有较好的实用性。     

基于FWADE-ELM的短时交通流预测方法

受道路环境和人为因素影响,实际交通系统可视为一个复杂的非线性动力系统,交通流数据具有较强的非线性、时变性和易受随机噪声影响等特征.针对复杂环境下的短时交通流预测问题,提出一种基于烟花差分进化混合算法-极限学习机的短时交通流预测方法.采用奇异谱分析方法滤除原始交通流数据中包含的噪声成分,降噪后的交通流数据用于训练极限学习...     

Application of temporal difference learning rules in short‐term traffic flow prediction

Studying dynamic behaviours of a transportation system requires the use of the system mathematical models as well as prediction of traffic flow in the system. Therefore, traffic flow prediction plays an important role in today's intelligent transportation systems. This article introduces a new approach to short‐term daily traffic flow prediction based on artificial neural networks. Among the family of neural networks, multi‐layer perceptron (MLP), radial basis function (RBF) neural network and wavenets have been selected as the three best candidates for performing traffic flow prediction. Moreover, back‐propagation (BP) has been adapted as the most efficient learning scheme in all the cases. It is shown that the coefficients produced by temporal signals improve the performance of the BP learning (BPL) algorithm. Temporal signals provide researchers with a new model of temporal difference BP learning algorithm (TDBPL). The capability and performance of TDBPL algorithm are examined by means of simulation in order to prove that the wavelet theory, with its multi‐resolution ability in comparison to RBF neural networks, is a suitable algorithm in traffic flow forecasting. It is also concluded that despite MLP applications, RBF neural networks do not provide negative forecasts. In addition, the local minimum problems are inevitable in MLP algorithms, while RBF neural networks and wavenet networks do not encounter them.     

社会情感算法优化神经网络的短时交通流预测

针对反向传播(BP)神经网络用于交通流预测易陷入局部最优且寻优速度慢的问题,采用了社会情感优化(SEO)BP神经网络的参数,以SEO中的个体为一个BP神经网络,以3种情绪为表现形式,通过个体间的合作竞争进行寻优.运用Levy、正态、柯西分布3种情绪随机选择策略,通过不同方式实现了以不同的概率选择不确定的情绪,使SEO中情绪更好地模拟人的正常心理变化.实验表明:该模型较其他模型更有利于搜寻全局最优解,能有效提高短时交通流的预测精度.     

Forecasting of short-term traffic-flow based on improved neurofuzzy models via emotional temporal difference learning algorithm

Bounded rationally idea, rather that optimization idea, have result and better performance in decision making theory. Bounded rationality is the idea in decision making, rationality of individuals is limited by the information they have, the cognitive limitations of their minds, and the finite amount of time they have to make decisions. The emotional theory is an important topic presented in this field. The new methods in the direction of purposeful forecasting issues, which are based on cognitive limitations, are presented in this study. The presented algorithms in this study are emphasizes to rectify the learning the peak points, to increase the forecasting accuracy, to decrease the computational time and comply the multi-object forecasting in the algorithms. The structure of the proposed algorithms is based on approximation of its current estimate according to previously learned estimates. The short term traffic flow forecasting is a real benchmark that has been studied in this area. Traffic flow is a good measure of traffic activity. The time-series data used for fitting the proposed models are obtained from a two lane street I-494 in Minnesota City, USA. The research discuss the strong points of new method based on neurofuzzy and limbic system structure such as Locally Linear Neurofuzzy network (LLNF) and Brain Emotional Learning Based Intelligent Controller (BELBIC) models against classical and other intelligent methods such as Radial Basis Function (RBF), Takagi–Sugeno (T–S) neurofuzzy, and Multi-Layer Perceptron (MLP), and the effect of noise on the performance of the models is also considered. Finally, findings confirmed the significance of structural brain modeling beyond the classical artificial neural networks.     

Recursive least-squares temporal difference learning for adaptive traffic signal control at intersection

This paper presents a new method to solve the scheduling problem of adaptive traffic signal control at intersection. The method involves recursive least-squares temporal difference (RLS-TD(λ)) learning that is integrated into approximate dynamic programming. The learning mechanism of RLS-TD(λ) is to make an adaptation of linear function approximation by updating its parameters based on environmental feedback. This study investigates the method implementation after modeling a traffic dynamic system at intersection in discrete time. In the model, different traffic control schemes regarding signal phase sequence are considered, especially the defined adaptive phase sequence (APS). By simulating traffic scenarios, RLS-TD(λ) is superior to TD(λ) for updating functional parameters in the approximation, and APS outperforms other conventional control schemes on reducing traffic delay. By comparing with other traffic signal control algorithms, the proposed algorithm yields satisfying results in terms of traffic delay and computation time.

     

基于改进GMDH算法的路口短时交通流量预测

城市交通是一个复杂的大系统,实时而准确的短时交通流量预测,可以为城市交通诱导和控制提供科学支持。针对GMDH算法建模泛化能力差的问题,结合集成学习的思想对GMDH算法进行改进,并将改进的算法应用到短时交通流量模型的构建中。结果表明,该方法可以有效地对短时交通流量进行预测,建模平均相对误差为1.10%,预测相对误差为0.58%。     

Exchange rate forecasting using a combined parametric and nonparametric self-organising modelling approach

Financial prediction has attracted a lot of interest due to the financial implications that the accurate prediction of financial markets can have. A variety of data driven modelling approaches have been applied but their performance has produced mixed results. In this study we apply both parametric (neural networks with active neurons) and nonparametric (analog complexing) self-organising modelling methods for the daily prediction of the exchange rate market. We also propose a combined approach where the parametric and nonparametric self-organising methods are combined sequentially, exploiting the advantages of the individual methods with the aim of improving their performance. The combined method is found to produce promising results and to outperform the individual methods when tested with two exchange rates: the American Dollar and the Deutche Mark against the British Pound.     

Recognizing machines with parametric and nonparametric learning methods using contextual information

Pattern recognition systems using information from the pattern which follows the present pattern are discussed. Parametric learning methods in the supervised and unsupervised machines are proposed and compare favorably with conventional methods. Furthermore, a semilinear machine with a nonparametric learning method is considered. The results of computer experiments with artificially generated data and with handprinted alphanumeric characters are given to show that the approach we adopt is quite useful for recognition of Markovian patterns.     

Short-term photovoltaic power forecasting using hybrid contrastive learning and temporal convolutional network under future meteorological information absence

Photovoltaic (PV) power generation is widely utilized to satisfy the increasing energy demand due to its cleanness and inexhaustibility. Accurate PV power forecasting can improve the penetration of PV power in the grid. However, it is pretty challenging to predict PV power in short-term under precious future meteorological information absence conditions. To address this problem, this study proposes the hybrid Contrastive Learning and Temporal Convolutional Network (CL-TCN), and this forecasting approach consists of two parts, including model training and adaptive processes of forecasting models. In the model training stage, this forecasting method firstly trains 18 TCN models for 18 time points from 9:00 a.m. to 17:30 p.m. These TCN models are trained by only using historical PV power data samples, and each model is used to predict the next half-hour power output. The adaptive process of models means that, in a practical forecasting stage, PV power samples from historical data are firstly evaluated and scored by a CL based data scoring mechanism to search for the most similar data samples to current measured samples. Then these similar samples are further applied to training a single above-mentioned well-trained TCN model to improve its performance in forecasting the next half-hour PV power. The experimental results tested at the time resolution of 30 min demonstrate that the proposed approach has superior performance in forecasting accuracy not only in smooth PV power samples but also in fluctuating PV power samples. Moreover, the proposed CL based data scoring mechanism can filter useless data samples effectively accelerating the forecasting process.     

Distributed learning automata based approach to inferring urban structure via traffic flow

Traffic flow can be used as a reference for knowledge generation, which is highly important in urban planning. One of the significant applications of traffic data is decision making about the structure of roads connecting zones of a city. It leads us to an optimal connection between important areas like business centers, shopping malls, construction sites, residential complexes, and other parts of a city which is the motivation of this research. The main question is how to infer the optimal connectivity network considering the current structure of an urban area and time-varying traffic dynamics. Therefore a novel formulation is created in this paper to solve the optimization problem using available data. A proposed algorithm is presented to infer the optimal structure that is a distributed learning automata-based approach. A matrix called estimated optimal connectivity represents the favorite structure and it is optimized utilizing signals about the current system and traffic dynamics from the environment. Two types of data, including synthetic and real-world, are used to show the algorithm’s ability. After many experiments, the algorithm showed capability of optimizing the structure by finding new paths connecting the most correlated areas.

     

SW-BiLSTM: a Spark-based weighted BiLSTM model for traffic flow forecasting

Accurate traffic flow forecasting (TFF) is significant for mitigating traffic congestion. To address the existing issues of calculation and storage in dealing with big traffic flow data using the traditional centralized models on a single machine, this paper presents a Spark-based Weighted Bidirectional Long Short-Term M emory (SW-BiLSTM) model to improve the robustness and accuracy of TFF. Specifically, the resilient distributed dataset (RDD) and the Kalman filter (KF) are utilized to preprocess large-scale trajectory data (e.g., GPS trajectories of taxicabs). Next, a distributed SW-BiLSTM model on Spark is put forward to enhance the accuracy and efficiency of TFF, combined with the normal distribution for weighing the influence degree of the interaction between adjacent road segments and the time window for implementing the optimization of BiLSTM. Finally, the experimental results on an empirical study with the real-world taxi GPS trajectory data indicate that, compared with ARIMA, LR, GNB, CNN, GRU, SAEs, BP, LSTM, and WND-LSTM (LSTM with a time window and a normal distribution), the MAPE value of SW-BiLSTM is decreased by 65.62%, 17.78%, 87.29%, 69.10%, 3.52%, 21.09%, 59.66%, 42.86%, and 1.22%, respectively. In particular, SW-BiLSTM is superior to BiLSTM with 15.83% accuracy improvement on average.

     

基于层次聚类和极限学习机的母线短期负荷预测

利用传统方法预测母线负荷时,通常选取离待测日相近的一段时间作为历史相似日进行模型训练,没有考虑其天气情况、星期类型、节假日等因素的影响,相似日与待测日特征相差较大。为解决以上问题,提出一种基于层次聚类（HC）和极限学习机（ELM）的母线负荷预测算法。首先使用层次聚类法将母线历史日负荷进行聚类,然后对层次聚类得出的聚类结果建立决策树,其次根据待测日的温度、湿度、星期和节假日类型等日属性在决策树中匹配出训练极限学习机预测模型的历史日负荷,最后建立极限学习机预测模型,对待测日母线日负荷进行预测。对两条不同母线的负荷进行了预测,与传统单一的极限学习机相比,所提算法的平均绝对百分比误差（MAPE）分别降低了1.4和0.8个百分点。实验结果表明,所提算法预测母线负荷具有更高的预测精度和稳定性。     

基于时间序列与BP-ANN的短时交通流速度预测模型研究

针对现有的交通流速度预测模型使用唯一数据集且模型单一的问题,提出一种时间序列与人工神经网络相结合的预测模型。该模型通过时间序列分别对实时数据和历史数据建模预测,并应用人工神经网络调整实时数据和历史数据的预测值。实验结果表明该预测模型能够将预测误差控制在7%以内,且能够对不同输入参数下的短时交通流速度进行有效预测。     

交通流量经验模态分解与神经网络短时预测方法

基于经验模态分解（EMD）和神经网络,提出了一种短时交通流量预测方法。通过EMD分解把交通流量分解成不同的模态,利用神经网络对分解后的各分量进行预测,再将预测值累加得到最终的预测结果。利用EMD与神经网络模型对I-800数据库实测交通流量数据进行预测,结果表明该方法具有很高的预测精度,明显优于直接采用神经网络的预测结果。