No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Origin-Destination Convolution Recurrent Network: A Novel OD Matrix Prediction Framework
Abstract
Origin-Destination(OD) Matrix Prediction is an important part of public transportation service which aims to predict the number of passenger demands from one region to another and capture the passengers' mobility patterns. This problem is challenging because it requires forecasting not only the number of demands within a region, but the origin and destination of each trip as well. To address this challenge, we propose an effective model, ODCRN(Origin-Destination Convolution Recurrent Network) which incorporates traffic context and bi-directional semantic information. First, we obtain the semantic embedded features of the region as the static traffic context by the Node2vec algorithm, and the traffic flow of the region is counted as the dynamic traffic context. Second, we construct two adjacency matrices which represent origin-destination and destination-origin travel demands within urban areas respectively based on the OD matrices of each time slot, and use the graph convolu-tional network to aggregate traffic context information of the semantic neighbors in both directions. Then, we use a unit constructed by GRU and the graph convolution network to capture the spatial-temporal correlations of the input data. Finally, we use those correlations and traffic contexts to predict the OD matrix for the next time slot. Our model is evaluated on TaxiNYC and TaxiCD datasets, and experimental results demonstrate the superiority of our ODCRN model against the state-of-the-art approaches.
... The learning rate is 0.0001, the batch size is 32, and the hidden layer feature dimension is 64. (7) ODCRN: OD Convolutional Recurrent Network [40]. ODCRN integrates recurrent and 2D graph convolutional neural networks to address the highly complex spatiotemporal dependencies in sequential OD matrices. ...
Short-term metro OD demand prediction during public health emergencies is a crucial task for the effective management and operation of metro systems. However, such emergencies tend to cause significant fluctuations in OD demand, making accurate prediction particularly challenging. To tackle this problem, this paper proposes a Multi-Frequency Spatial-Temporal Graph Neural Network (MFST-GNN) to accurately predict the metro OD demand during public health emergencies. Specifically, multiple OD demand patterns, including real-time, daily, and weekly OD demand are leveraged to extract the periodicity spatial-temporal features of OD demand. A novel multi-frequency temporal feature extraction module is developed to capture the periodic temporal features, while an adaptive spatial feature extraction module is introduced to learn the complex hidden spatial features. Moreover, event-related information is collected and integrated into the OD features to study the impact of events on OD demand. The effectiveness of the proposed model is validated by a large-scale real-world metro OD dataset, with comparative analysis against benchmark prediction models. Results demonstrate its superior performance and practical application potential.
Forecasting the flow of crowds is of great importance to traffic management and public safety, yet a very challenging task affected by many complex factors, such as inter-region traffic, events and weather. In this paper, we propose a deep-learning-based approach, called ST-ResNet, to collectively forecast the in-flow and out-flow of crowds in each and every region through a city. We design an end-to-end structure of ST-ResNet based on unique properties of spatio-temporal data. More specifically, we employ the framework of the residual neural networks to model the temporal closeness, period, and trend properties of the crowd traffic, respectively. For each property, we design a branch of residual convolutional units, each of which models the spatial properties of the crowd traffic. ST-ResNet learns to dynamically aggregate the output of the three residual neural networks based on data, assigning different weights to different branches and regions. The aggregation is further combined with external factors, such as weather and day of the week, to predict the final traffic of crowds in each and every region. We evaluate ST-ResNet based on two types of crowd flows in Beijing and NYC, finding that its performance exceeds six well-know methods.
A dynamic origin-destination demand estimation model for planning applications with real-time link counts from multiple days is presented. Based on an iterative bilevel estimation framework, the upper-level problem is to minimize both the deviation between estimated link flows and real-time link counts and the deviation between estimated time-dependent demand and given historical static demand. These two types of deviations are combined into a weighted objective function, where the weighting value is determined by an interactive approach to obtain the best compromise solution. The single-day formulation is further extended to use link counts from multiple days to estimate the variation in traffic demand over multiple days. A case study based on the Irvine test bed network is conducted to illustrate the methodology and estimate day-to-day demand patterns. The application illustrates considerable benefits in analyzing the demand dynamics with multiday data.
The availability of user check-in data in large volume from the rapid growing location based social networks (LBSNs) enables many important location-aware services to users. Point-of-interest (POI) recommendation is one of such services, which is to recommend places where users have not visited before. Several techniques have been recently proposed for the recommendation service. However, no existing work has considered the temporal information for POI recommendations in LBSNs. We believe that time plays an important role in POI recommendations because most users tend to visit different places at different time in a day, \eg visiting a restaurant at noon and visiting a bar at night. In this paper, we define a new problem, namely, the time-aware POI recommendation, to recommend POIs for a given user at a specified time in a day. To solve the problem, we develop a collaborative recommendation model that is able to incorporate temporal information. Moreover, based on the observation that users tend to visit nearby POIs, we further enhance the recommendation model by considering geographical information. Our experimental results on two real-world datasets show that the proposed approach outperforms the state-of-the-art POI recommendation methods substantially.
Many problems in transport planning and management tasks require an origindestination (O-D) matrix to represent the travel pattern. However, O-D matrices obtained through a large scale survey such as home or roadside interviews, tend to be costly, labour intensive and time disruptive to trip makers. Therefore, the use of low cost and easily available data is particularly attractive.
The need of low-cost methods to estimate current and future O-D matrices is even more valuable in developing countries because of the rapid changes in population, economic activity and land use. Models of transport demand have been used for many years to synthesize O-D matrices in study areas. A typical example of this is the gravity model; its functional form, plus the appropriate values for the parameters involved, is employed to produce acceptable matrices representing trip making behaviour for many trip purposes and time periods.
The work reported in this paper has combined the advantages of acceptable travel demand models with the low cost and availability of traffic counts. Three types of demand models have been used: gravity (GR), opportunity (OP) and gravity-opportunity (GO) models. Three estimation methods have been developed to calibrate these models from traffic counts, namely: non-linear-least-squares (NLLS), weighted-non-linear-least-squares (WNLLS) and maximumlikelihood (ML).
The 1978 Ripon (urban vehicle movement) survey was used to test these methods. They were found to perform satisfactorily since each calibrated model reproduced the observed O-D matrix fairly closely. The tests were carried out using two assignment techniques, all-or-nothing and the stochastic method due to Burrell, in determining the routes taken through the network.
Ride-hailing platforms generally provide various service options to customers, such as solo ride services, shared ride services, etc. It is generally expected that demands for different service modes are correlated, and the prediction of demand for one service mode can benefit from historical observations of demands for other service modes. Moreover, an accurate joint prediction of demands for multiple service modes can help the platforms better allocate and dispatch vehicle resources. Although there is a large stream of literature on ride-hailing demand predictions for one specific service mode, few efforts have been paid towards joint predictions of ride-hailing demands for multiple service modes. To address this issue, we propose a deep multi-task multi-graph learning approach, which combines two components: (1) multiple multi-graph convolutional (MGC) networks for predicting demands for different service modes, and (2) multi-task learning modules that enable knowledge sharing across multiple MGC networks. More specifically, two multi-task learning structures are established. The first one is the regularized cross-task learning, which builds cross-task connections among the inputs and outputs of multiple MGC networks. The second one is the multi-linear relationship learning, which imposes a prior tensor normal distribution on the weights of various MGC networks. Although there are no concrete bridges between different MGC networks, the weights of these networks are constrained by each other and subject to a common prior distribution. Evaluated with the for-hire-vehicle datasets in Manhattan, we show that our proposed approach outperforms the benchmark algorithms in prediction accuracy for different ride-hailing modes.
With the rapid development of mobile-internet technologies, on-demand ride-sourcing services have become increasingly popular and largely reshaped the way people travel. Demand prediction is one of the most fundamental components in supply-demand management systems of ride-sourcing platforms. With an accurate short-term prediction for origin-destination (OD) demand, the platforms make precise and timely decisions on real-time matching, idle vehicle reallocations, and ride-sharing vehicle routing, etc. Compared to the zone-based demand prediction that has been examined in many previous studies, OD-based demand prediction is more challenging. This is mainly due to the complicated spatial and temporal dependencies among the demand of different OD pairs. To overcome this challenge, we propose the Spatio-Temporal Encoder-Decoder Residual Multi-Graph Convolutional network (ST-ED-RMGC), a novel deep learning model for predicting ride-sourcing demand of various OD pairs. Firstly, the model constructs OD graphs, which utilize adjacent matrices to characterize the non-Euclidean pair-wise geographical and semantic correlations among different OD pairs. Secondly, based on the constructed graphs, a residual multi-graph convolutional (RMGC) network is designed to encode the contextual-aware spatial dependencies, and a long-short term memory (LSTM) network is used to encode the temporal dependencies, into a dense vector space. Finally, we reuse the RMGC networks to decode the compressed vector back to OD graphs and predict the future OD demand. Through extensive experiments on the for-hire-vehicles datasets in Manhattan, New York City, we show that our proposed deep learning framework outperforms the state-of-arts by a significant margin.
Due to the widespread applications in real-world scenarios, metro ridership prediction is a crucial but challenging task in intelligent transportation systems. However, conventional methods either ignore the topological information of metro systems or directly learn on physical topology, and cannot fully explore the patterns of ridership evolution. To address this problem, we model a metro system as graphs with various topologies and propose a unified Physical-Virtual Collaboration Graph Network (PVCGN), which can effectively learn the complex ridership patterns from the tailor-designed graphs. Specifically, a physical graph is directly built based on the realistic topology of the studied metro system, while a similarity graph and a correlation graph are built with virtual topologies under the guidance of the inter-station passenger flow similarity and correlation. These complementary graphs are incorporated into a Graph Convolution Gated Recurrent Unit (GC-GRU) for spatial-temporal representation learning. Further, a Fully-Connected Gated Recurrent Unit (FC-GRU) is also applied to capture the global evolution tendency. Finally, we develop a Seq2Seq model with GC-GRU and FC-GRU to forecast the future metro ridership sequentially. Extensive experiments on two large-scale benchmarks (e.g., Shanghai Metro and Hangzhou Metro) well demonstrate the superiority of our PVCGN for station-level metro ridership prediction. Moreover, we apply the proposed PVCGN to address the online origin-destination (OD) ridership prediction and the experiment results show the universality of our method. Our code and benchmarks are available at
https://github.com/HCPLab-SYSU/PVCGN
.
Ride-hailing applications are becoming more and more popular for providing drivers and passengers with convenient ride services, especially in metropolises like Beijing or New York. To obtain the passengers' mobility patterns, the online platforms of ride services need to predict the number of passenger demands from one region to another in advance. We formulate this problem as an Origin-Destination Matrix Prediction (ODMP) problem. Though this problem is essential to large-scale providers of ride services for helping them make decisions and some providers have already put it forward in public, existing studies have not solved this problem well. One of the main reasons is that the ODMP problem is more challenging than the common demand prediction. Besides the number of demands in a region, it also requires the model to predict the destinations of them. In addition, data sparsity is a severe issue. To solve the problem effectively, we propose a unified model, Grid-Embedding based Multi-task Learning (GEML) which consists of two components focusing on spatial and temporal information respectively. The Grid-Embedding part is designed to model the spatial mobility patterns of passengers and neighboring relationships of different areas, the pre-weighted aggregator of which aims to sense the sparsity and range of data. The Multi-task Learning framework focuses on modeling temporal attributes and capturing several objectives of the ODMP problem. The evaluation of our model is conducted on real operational datasets from UCAR and Didi. The experimental results demonstrate the superiority of our GEML against the state-of-the-art approaches.
Taxi demand prediction has recently attracted increasing research interest due to its huge potential application in large-scale intelligent transportation systems. However, most of the previous methods only considered the taxi demand prediction in origin regions, but neglected the modeling of the specific situation of the destination passengers. We believe it is suboptimal to preallocate the taxi into each region-based solely on the taxi origin demand. In this paper, we present a challenging and worth-exploring task, called taxi origin-destination demand prediction, which aims at predicting the taxi demand between all-region pairs in a future time interval. Its main challenges come from how to effectively capture the diverse contextual information to learn the demand patterns. We address this problem with a novel contextualized spatial-temporal network (CSTN), which consists of three components for the modeling of local spatial context (LSC), temporal evolution context (TEC), and global correlation context (GCC), respectively. First, an LSC module utilizes two convolution neural networks to learn the local spatial dependencies of taxi, demand respectively, from the origin view and the destination view. Second, a TEC module incorporates the local spatial features of taxi demand and the meteorological information to a Convolutional Long Short-term Memory Network (ConvLSTM) for the analysis of taxi demand evolution. Finally, a GCC module is applied to model the correlation between all regions by computing a global correlation feature as a weighted sum of all regional features, with the weights being calculated as the similarity between the corresponding region pairs. The extensive experiments and evaluations on a large-scale dataset well demonstrate the superiority of our CSTN over other compared methods for the taxi origin-destination demand prediction.
This paper proposes a deep learning based multitask learning (MTL) model to predict network-wide traffic speed, and two methods to improve its prediction performance. The nonlinear Granger causality analysis detects the spatiotemporal causal relationships among various links to select the most informative features for the MTL model. Bayesian optimization optimizes the hyperparameters of the MTL model with limited computational costs. Then, numerical experiments are conducted with taxis’ GPS data in an urban road network of Changsha, China, and some conclusions are drawn as follows. The deep learning based MTL model outperforms four deep learning based single task learning (STL) models (i.e., Gated Recurrent Units network, Long Short-term Memory network, Convolutional Gated Recurrent Units network and Temporal Convolutional Network) and three other classic models (i.e., Support Vector Machine, k-Nearest Neighbors and Evolving Fuzzy Neural Network), due to its better capability of digging the spatiotemporal patterns and capturing the nonlinear mapping relationships at a network level. The nonlinear Granger causality test provides a reliable guide to select the informative features from network-wide links for the MTL model. Compared with two other optimization approaches (i.e., grid search and random search), Bayesian optimization yields a better tuning performance for the MTL model in the prediction accuracy under the budgeted computation cost. In summary, the deep learning based MTL model with nonlinear Granger causality analysis and Bayesian optimization promises the accurate and efficient traffic prediction for a large-scale network.
Predicting flows (e.g. the traffic of vehicles, crowds and bikes), consisting of the in-out traffic at a node and transitions between different nodes, in a spatio-temporal network plays an important role in transportation systems. However, this is a very challenging problem, affected by multiple complex factors, such as spatial correlations between different locations, temporal correlations among different time intervals, and external factors (like events and weather). In addition, the flow at a node (called node flow) and transitions between nodes (edge flow) mutually influence each other. To address these issues, we propose a multitask deep-learning framework that simultaneously predicts the node flow and edge flow throughout a spatio-temporal network. Using fully convolutional networks, our approach designs two sophisticated models for predicting node flow and edge flow respectively. Two models are connected by coupling their latent representations of middle layers, and trained together. The external factors are also integrated into the framework through a gating fusion mechanism. In the edge flow prediction model, we employ an embedding component to deal with the sparse transitions between nodes. We evaluate our method based on the taxicab data in Beijing and New York City.Experimental results show advantages of our method beyond 11 baselines, such as ConvLSTM, CNN, and Markov Random Field.
Urban mass transit is playing a significant role in supporting and promoting urban development. An important indicator for the planning and design of urban rail transit may be succinctly summarized by passenger flow models within a peak hour; one important feature of the model is the maximum single-direction flow. To determine this feature, it is necessary to forecast passengers' departure time and route choice during a peak period. As the basis of this process, the peak-period station-to-station origin-destination (OD) matrix reflects passengers' travel needs. This paper tests the traditional gravity models to find the pattern that forecasts the peak-period station-to-station OD matrix in urban rail transit. A real-world case study of Chongqing, China, is used as a model performance measure. To alleviate its over-estimation when the effect of the deterrence function between two stations is too small, the gravity-model-based peak period coefficient (PPC) model is introduced. By comparing the PPC and gravity models using the same dataset, the results indicate that the PPC model is superior to the gravity model. The standard deviation of the PPC model is 12.90 passengers, which is 56.02% lower than that of the gravity model, which is 29.33 passengers. © 2018, Editorial Department of Journal of Tongji University. All right reserved.
Prediction tasks over nodes and edges in networks require careful effort in engineering features used by learning algorithms. Recent research in the broader field of representation learning has led to significant progress in automating prediction by learning the features themselves. However, present feature learning approaches are not expressive enough to capture the diversity of connectivity patterns observed in networks. Here we propose node2vec, an algorithmic framework for learning continuous feature representations for nodes in networks. In node2vec, we learn a mapping of nodes to a low-dimensional space of features that maximizes the likelihood of preserving network neighborhoods of nodes. We define a flexible notion of a node's network neighborhood and design a biased random walk procedure, which efficiently explores diverse neighborhoods. Our algorithm generalizes prior work which is based on rigid notions of network neighborhoods, and we argue that the added flexibility in exploring neighborhoods is the key to learning richer representations.
We demonstrate the efficacy of node2vec over existing state-of-the-art techniques on multi-label classification and link prediction in several real-world networks from diverse domains. Taken together, our work represents a new way for efficiently learning state-of-the-art task-independent representations in complex networks.
Estimation of origin–destination (O–D) matrices from link count data is considered. This problem is challenging because the number of parameters to be estimated is typically larger than the number of network links. As a result, it is (usually) impossible to identify a unique optimal estimate of the O–D matrix from mean link traffic counts. However, information from the covariance matrix of link count data collected over a sequence of days can relieve this problem of indeterminacy. This fact is illustrated through a simple example. The use of second-order statistical properties of the data in O–D matrix estimation is then explored, and a class of estimators proposed. Practical problems of model mis-specification are discussed and some avenues for future research outlined.
This paper presents an in-depth study of the methodology for estimating or updating origin-to-destination trip matrices from traffic counts. Following an analysis of the statistical foundation of the estimation and updating problems, various basic approaches are reviewed using a generic traffic assignment map. Computational issues related to specific assignment maps and estimation models for both road and transit networks are then discussed. Finally, additional insight into the relative performance of several estimators is provided by a set of test problems with varying input data.
Many underlying relationships among data in several areas of science and engineering, e.g., computer vision, molecular chemistry, molecular biology, pattern recognition, and data mining, can be represented in terms of graphs. In this paper, we propose a new neural network model, called graph neural network (GNN) model, that extends existing neural network methods for processing the data represented in graph domains. This GNN model, which can directly process most of the practically useful types of graphs, e.g., acyclic, cyclic, directed, and undirected, implements a function tau(G,n) isin IR
m
that maps a graph G and one of its nodes
n
into an
m
-dimensional Euclidean space. A supervised learning algorithm is derived to estimate the parameters of the proposed GNN model. The computational cost of the proposed algorithm is also considered. Some experimental results are shown to validate the proposed learning algorithm, and to demonstrate its generalization capabilities.
This paper proposes a dynamic origin-destination (OD) estimation method to extract valuable point-to-point split-fraction information from automatic vehicle identification (AVI) counts without estimating market-penetration rates and identification rates of AVI tags. A nonlinear ordinary least-squares estimation model is presented to combine AVI counts, link counts, and historical demand information into a multiobjective optimization framework. A joint estimation formulation and a one-sided linear-penalty formulation are further developed to take into account possible identification and representativeness errors, and the resulting optimization problems are solved by using an iterative bilevel estimation procedure. Based on a synthetic data set, this study shows the effectiveness of the proposed estimation models under different market-penetration rates and identification rates
The development of maximumentropy (me) and bayesian-inference (bi) estimation methods for calibrating transport demand models based on link volume information
- O Z Tamin
- H Hidayat
- A K Indriastuti
O. Z. Tamin, H. Hidayat, and A. K. Indriastuti, "The development of maximumentropy (me) and bayesian-inference (bi) estimation methods for calibrating transport demand models based on link volume information," in EASTS, vol. 4, 2003,
pp. 630-647.
Diffusion convolutional recurrent neural network: Data-driven traffic forecasting
- Yaguang Li
Li, Yaguang, et al. "Diffusion convolutional recurrent neural network: Data-driven
traffic forecasting." arXiv preprint arXiv:1707.01926 (2017).