Showing papers by "Zhiyuan Liu published in 2021"

A multi-stage stochastic optimization approach to the stop-skipping and bus lane reservation schemes

Abstract: The bus stop-skipping scheme is designed to reduce the number of stops and increase bus operating speed. But in a mixed traffic environment, buses can hardly maintain a high and stable operating sp...

Exploring a large-scale multi-modal transportation recommendation system

Abstract: The emergence of navigation applications with multi-modal trip planning services has brought about the demand for the multi-modal transportation recommendation systems. In this paper, we explore the problem of large-scale multi-modal transportation recommendation and propose a novel travel mode recommendation system for a multi-modal transportation system. In the proposed model, the feature engineering focuses on the application scenario of the multi-modal transportation recommendation, and is designed from multiple perspectives of users, travel modes, locations, and time. To learn a better representation of the co-occurrence, we construct a bipartite graph for the Origin-Destination (OD) pair and the User-OD pair of all the query records then transformed nodes in the bipartite graph to feature vectors using a graph-embedding technique. Finally, we propose a post-processing technique to handle the inconsistency between the objective function and evaluation metric. Experimental results from a city-wide multi-modal transportation recommendation indicate that our proposed model is superior to the existing method of navigation service providers.

An s-shaped three-parameter (S3) traffic stream model with consistent car following relationship

Abstract: In this study, a new s-shaped three-parameter (S3) traffic flow model is proposed to represent the relationships between three fundamental variables (i.e., flow, speed, and density) in highway traffic. An s-shaped speed-density function is proposed to capture the speed-density relationship under a wide range of possible densities. A consistent car-following model was derived in terms of the proposed s-shaped speed-density function. Both the S3 macroscopic model and the derived microscopic car-following model were calibrated using real-world traffic data.

A systematic review of prediction methods for emergency management

Abstract: With the trend of global warming and destructive human activities, the frequent occurrences of catastrophes have posed devastating threats to human life and social stability worldwide. The emergency management (EM) system plays a significant role in saving people's lives and reducing property damage. The prediction system for the occurrence of emergency events and resulting impacts is widely recognized as the first stage of the EM system, the accuracy of which has a significant impact on the efficiency of resource allocation, dispatching, and evacuation. In fact, the number and variety of contributions to prediction techniques, such as statistic analysis, artificial intelligence, and simulation method, are exploded in recent years, motivating the need for a systematic analysis of the current works on disaster prediction. To this end, this paper presents a systematic review of contributions on prediction methods for emergency occurrence and resource demand of both natural and man-made disasters. Through a detailed discussion on the features of each type of emergency event, this paper presents a comprehensive survey of state-of-the-art prediction technologies which have been widely applied in EM. After that, we summarize the challenges of current efforts and point out future directions.

Building Personalized Transportation Model for Online Taxi-Hailing Demand Prediction

Abstract: The accurate prediction of online taxi-hailing demand is challenging but of significant value in the development of the intelligent transportation system. This article focuses on large-scale online taxi-hailing demand prediction and proposes a personalized demand prediction model. A model with two attention blocks is proposed to capture both spatial and temporal perspectives. We also explored the impact of network architecture on taxi-hailing demand prediction accuracy. The proposed method is universal in the sense that it is applicable to problems associated with large-scale spatiotemporal prediction. The experimental results on city-wide online taxi-hailing demand dataset demonstrate that the proposed personalized demand prediction model achieves superior prediction accuracy.

Automatic Feature Engineering for Bus Passenger Flow Prediction Based on Modular Convolutional Neural Network

Abstract: Deep Neural Network (DNN) has been applied in a wide range of fields due to its exceptional predictive power. In this paper, we explore how to use DNN to solve the large-scale bus passenger flow prediction problem. Currently, most existing methods designed for the passenger flow prediction problem are based on a single view, which is insufficient to capture the dynamics in passenger flow fluctuation. Thus, we analyze the passenger flow from scopes on both macroscopic and microscopic levels, in order to take full advantage of the information from a variety of views. To better understand the role of different views, decision-tree-based models are used in modeling and predicting passenger flow. The defects and key features of decision-tree-based models are then analyzed. The results of the analysis can assist the architecture design of the deep learning network. Inspired by the feature engineering of decision-tree-based models, a modular convolutional neural network is designed, which contains automatic feature extraction block, feature importance block, fully-connected block, and data fusion block. The proposed model is evaluated on the city-wide public transport datasets in Nanjing, China, involving 1,091 bus lines in total. The experiment results demonstrate the outstanding performance of the proposed method in real situations.

Estimation of urban network capacity with second-best constraints for multimodal transport systems

Abstract: Transport network capacity enhancement is a significant aspect of urban transport planning and demand management, and a suitable measurement of the network capacity is of considerable importance. In this paper, the network capacity with second-best constraints (NCSC) is investigated to meet some specific development requirements of urban transport networks. Herein, the network capacity is restricted to an inferior “second-best solution”, due to various concerns/constraints regarding the public transport mode share, serviceability, and emissions, etc. For the sake of presentation, these constraints are termed as second-best constraints, and the NCSC problem can also be referred as second-best network capacity (SBNC) problem. A bi-level model is formulated to analyse the NCSC problem. The upper-level model maximizes the total origin-destination (OD) demand, which incorporates the second-best constraints into consideration. The lower-level model is a transport network equilibrium model, which measures the network performance under a given OD demand pattern. To better investigate some important second-best constraints (e.g., public transport mode share) and also the demand elasticity, the modelling framework is extended to a multimodal transport network. An exact solution method is developed for the NCSC problem; wherein, a modified improved gradient projection (MIGP) algorithm is designed for the lower-level multimodal flow equilibrium problem, and a tailored sensitivity analysis-based (SAB) method is employed for solving the NCSC problem. The proposed models and solution methods are verified by numerical examples, demonstrating that NCSC can be an efficient tool for transport planning and management.

An agent-based model for real-time bus stop-skipping and holding schemes

Abstract: This paper develops a real-time agent-based simulation model to optimize the dynamic bus stop-skipping and holding schemes. The proposed agent-based simulation model is capable of modeling, schedul...

Behavior2vector: Embedding Users' Personalized Travel Behavior to Vector

Abstract: We investigate how to effectively and efficiently embed users' personalized travel behaviors to vectors in this paper. Based on an example scenario of travel mode choice in intelligent transportation system, three data structures representing users' travel behaviors are defined, namely heterogeneous graph of users' travel behaviors, user travel behavior k-partite graph, and personalized user travel behavior sentence set. This paper systematically analyzes the principle of existing methods and provides intuitions for the problem of learning travel behavior representation in intelligent transportation system. Then we propose the Behavior2vector, which is an improved method tailored for embedding users' personalized travel behaviors to vectors. In our experiments, we design a travel mode choice model based on machine learning, which uses both hand-crafted basic features and embedded vector features. We further quantify the impact of various factors on travel mode choice and use travel big data to test the hypothesis of traffic assignment models, e.g., travelers always choose the path with the shortest path. In addition, we also compared with the existing graph embedding methods and essentially discussed their advantages and disadvantages.

The modulation of attentional deployment on regret: an event-related potential study.

Abstract: Adopting a sequential risk-taking task, this study explored the modulation of attentional deployment on regret. Attentional deployment was manipulated during outcome feedback of the task by highlighting different parts to induce participants to focus on collected gains (GF context) or missed chances (MF context). The control context without attentional deployment manipulation was also set. Behaviorally, compared to the control context, participants felt less regret in the GF context but more regret in the MF context. Event-related potential results showed that the GF context elicited stronger reward positivity and late positive potential (LPP) than the control context. Furthermore, openness (NEO Five-Factor Inventory) negatively predicted the amplitude of LPP in the GF context. Source localization indicated that the superior frontal gyrus showed stronger activation in the GF context than in the control context during the time window of LPP. These results suggested that focusing on collected gains was an effective way to repress regret and that the LPP component played a key role in this process.

Modelling the energy consumption of electric vehicles under uncertain and small data conditions

Abstract: This study models the energy consumption of electric vehicles (EVs) under uncertain and small data conditions by combining the machine learning method and the idea of controlled experiments. We propose a Machine Learning-Control Variable model, termed the MLCV model, to estimate the trip energy consumption of EVs. Different data augmentation methods, ensemble methods, sampling factors are adopted as the parameters of the proposed method. Through parameter search, the accuracy of the base learner can be further improved. Our method utilizes real driving behaviours that are generated by real drivers and collected in a complex urban environment, making the approach generalizable. The experimental results demonstrate that the proposed MLCV model is superior to existing machine learning models in terms of estimation accuracy.

An improved learning-and-optimization train fare design method for addressing commuting congestion at CBD stations

Abstract: This study proposes an improved learning-and-optimization train fare design method to deal with the commuting congestion of train stations at the central business district (CBD). The conventional learning-and-optimization scheme needs accurate boarding/alighting demand to update the train fare in each trial. However, when congestion happens, the observed boarding/alighting demand will be larger than the actual boarding/alighting demand due to the delays and the longer dwelling time. Thus, the actual boarding/alighting demand is not available in practice. The improved algorithm deals with this issue by using inexact and less information to determine the new trial fare during the iteration. Namely, the improved method bypasses the conditions that may lead to biased results so as to significantly enhance the reliability of the learning-and-optimization method. The simplified algorithm also makes this method more practical. The convergence property of the proposed algorithm is rigorously proved and the convergence rate is demonstrated to be exponential. Numerical studies are performed to demonstrate the efficiency of the improved learning-and-optimization method.

Bus OD matrix reconstruction based on clustering Wi-Fi probe data

Abstract: The estimation of citywide passenger demand plays a vital role in system planning, operation, and management of the urban transit system. The Wi-Fi probe data, one of the emerging crowdsourcing dat...

Optimal Congestion Pricing with Day-to-Day Evolutionary Flow Dynamics: A Mean–Variance Optimization Approach

Abstract: This paper investigates the optimal congestion pricing problem that considers day-to-day evolutionary flow dynamics. Under the circumstance that traffic flows evolve from day to day and the system might be in a non-equilibrium state during a certain period of days after implementing (or adjusting) a congestion toll scheme, it is questionable to use an equilibrium-based index under steady state as the objective to measure the performance of a congestion toll scheme. To this end, this paper proposes a mean–variance-based congestion pricing scheme, which is a robust optimization model, to consider the evolution process of traffic flow dynamics in the optimal toll design problem. More specifically, in the mean–variance-based toll scheme, travelers aim to minimize the variance of expected total travel costs (ETTCs) on different days to reduce risk in daily travels, while the average ETTC over the whole planning period is restricted to being no larger than a predetermined target value set by the authorities. A metaheuristic approach based on the whale optimization algorithm is designed to solve the proposed mean–variance-based day-to-day dynamic congestion pricing problem. Finally, a numerical experiment is conducted to validate the effectiveness of the proposed model and solution algorithm. Results show that the used 9-node network can reach a steady state within 18 days after implementing the mean–variance-based congestion pricing, and the optimal toll scheme can be also obtained with this toll strategy.

An improved parallel block coordinate descent method for the distributed computing of traffic assignment problem

Abstract: An improved parallel block coordinate descent (iPBCD) algorithm for solving the user equilibrium traffic assignment problem is proposed. The iPBCD algorithm is developed based on the parallel block...

Semi-Analytical Path Delay Variation Model With Adjacent Gates Decorrelation for Subthreshold Circuits

Abstract: The subthreshold circuit is a practical design style for the ultralow-power applications, but its timing estimation is a challenge due to the increasing local variation effects. The delay variation of adjacent gates is not independent because of input slew variation caused by the precedent gate, so their correlation effects are difficult to model and estimate. This article proposes a semi-analytical statistical delay model considering local variation for the subthreshold region, that is, the combination of analytical and simulation-based method. First, it decorrelates the slew influence between adjacent stages by dividing delay and output slew model into fast/slow input cases and dividing delay variation model into process variation and input slew variation. Then, it can be applied into multi-PVT conditions with a one-time SPICE nominal simulation by analyzing the independence of variability and relative variability of step input gate delay variance with output load capacitance and process, voltage, and temperature (PVT). Finally, experiments are carried out for different benchmarks, processes, voltages, and temperatures (BPVTs). The average errors of variance on different BPVTs are 4.8%, 3.1%, 4.0%, and 4.7%. Compared with other analytical works, the accuracies’ improvements of three metrics (variance, variability, and max delay) are $8.3\times $ , $9.6\times $ , and $2.7\times $ by the mean error at all test benchmarks. Compared with industrial method LVF, it has a comparable error in max path delay and runtime, and less three orders of magnitudes than LVF in characterization time and stored data (from TB to GB) at all test benchmarks.

Improved Dynamic Time Warping Algorithm for Bus Route Trajectory Curve Fitting

Abstract: Applying the electronic route map is a cornerstone for improving the management and operational efficiency of transportation systems. However, the practice of electronic route map designed ...

Joint optimisation of transfer location and capacity for a capacitated multimodal transport network with elastic demand: a bi-level programming model and paradoxes

Abstract: With the growing attention toward developing a multimodal transport system to enhance urban mobility, there is an increasing need to construct new infrastructures, rebuild or expand the existing ones, to accommodate the current and newly generated travel demand. Therefore, this study develops a bi-level model that simultaneously determines the location and capacity of the transfer infrastructure to be built considering the elastic demand in a multimodal transport network. The upper-level problem is formulated as a mixed-integer linear programming problem, whereas the lower-level problem is a combined trip distribution/modal split/assignment model that depicts both the destination and route choices of travellers via a multinomial logit model. Numerical studies are conducted to demonstrate the occurrence of two Braess-like paradox phenomena in a multimodal transport network. The first one states that under fixed demand, constructing new parking spaces to provide the usage of park-and-ride services could deteriorate the system performance measured by the total passengers’ travel time, while the second one reveals that under elastic demand, increasing the parking capacity for park-and-ride services to promote its usage may fail, which would be represented by the decline in their modal share. Meanwhile, a numerical example also suggests that constructing transfer infrastructures at distributed stations outperforms building a large transfer centre in terms of attracting travellers using sustainable transit modes.

Pre-stimulus alpha activities predict the confidence in subjective judgment by modulating post-stimulus theta oscillation

Abstract: Even when making arbitrary decisions, people tend to feel varying levels of confidence, which is associated with the pre-stimulus neural oscillation of the brain. We investigated varying confidence in a pure subjective judgment task, and how this confidence was predicted by pre-stimulus alpha oscillations. Participants made pure subjective judgments where their prior experience seems to be helpful but actually useless, and their fluctuating confidence was related to the choice boundary process rather than the evidence accumulation process, suggesting participants underwent varying confidence resulting from the internal signals. With EEG and MEG analyses, we not only revealed the linkage between confidence and pre-stimulus alpha activities, but also successfully located this linkage onto decision-making relevant brain areas, i.e. MCC/PCC and SMA. Moreover, we unveiled a specific pathway underlying such linkage, that is, the influence of pre-stimulus alpha activities on decision confidence was fulfilled through modulating post-stimulus theta activities of SMA.