A stochastic optimal control approach for real-time traffic routing considering demand uncertainties and travelers’ choice heterogeneity

Pi, Xidong; Qian, Sean

doi:10.1016/j.trb.2017.06.002

Cited by 19 publications

(4 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recently, studies have examined the optimization of traffic signal timings as an active management tool for nonrecurring congestion (3)(4)(5)(6). Traffic assignment models, equipped with behavior models that characterize travelers' behavior changes after incidents, and predictioncorrection models are often built to simulate the timedependent diverted traffic flow under pre-defined incident scenarios (7)(8)(9)(10)(11)(12)(13). The signal timings, optimized for a given incident scenario, then favor specified directional movements to minimize the induced congestion (14).…”

Section: Traffic Signal Timings For Incident Managementmentioning

confidence: 99%

Learning to Recommend Signal Plans under Incidents with Real-Time Traffic Prediction

Yao

Qian

2020

Transportation Research Record

View full text Add to dashboard Cite

The main question to address in this paper is to recommend optimal signal timing plans in real time under incidents by incorporating domain knowledge developed with the traffic signal timing plans tuned for possible incidents, and learning from historical data of both traffic and implemented signals timing. The effectiveness of traffic incident management is often limited by the late response time and excessive workload of traffic operators. This paper proposes a novel decision-making framework that learns from both data and domain knowledge to real-time recommend contingency signal plans that accommodate non-recurrent traffic, with the outputs from real-time traffic prediction at least 30 minutes in advance. Specifically, considering the rare occurrences of engagement of contingency signal plans for incidents, we propose to decompose the end-to-end recommendation task into two hierarchical models -real-time traffic prediction and plan association. We learn the connections between the two models through metric learning, which reinforces partial-order preferences observed from historical signal engagement records. We demonstrate the effectiveness of our approach by testing this framework on the traffic network in Cranberry Township in 2019. Results show that our recommendation system has a precision score of 96.75% and recall of 87.5% on the testing plan, and make recommendation of an average of 22.5 minutes lead time ahead of Waze alerts. The results suggest that our framework is capable of giving traffic operators a significant time window to access the conditions and respond appropriately.

show abstract

Section: Traffic Signal Timings For Incident Managementmentioning

confidence: 99%

Learning to Recommend Signal Plans under Incidents with Real-Time Traffic Prediction

Yao

Qian

2020

Transportation Research Record

View full text Add to dashboard Cite

show abstract

“…Note the vehicles controlled in each time t can be different. The NDL D F rs (t) is estimated by RV data based on Equation 13 or Equation 43, and its estimation is usually delayed, meaning that D F rs (t) is not immediately available at time t. It may be available after lag amount of time intervals. To address this issue, any time series prediction methods can be adopted to obtain a forecast of D F rs (t) .…”

Section: Traffic Management Through User Optimal Routingmentioning

confidence: 99%

“…In a realworld traffic management problem, controlling all the vehicles in the network is implausible. Instead, traffic delay can be considerably reduced by controlling a relatively small portion of vehicles [43,62,51].…”

Section: Traffic Management Through User Optimal Routingmentioning

confidence: 99%

Measuring and reducing the disequilibrium levels of dynamic networks with ride-sourcing vehicle data

Qian

2020

Transportation Research Part C: Emerging Technologies

View full text Add to dashboard Cite

“…A key big-data insight from this type of data is the repeated observations over a period of time that allows for a richer characterization of uncertainty and the mobility-demand processes. Some of the studies in real-time routing utilize the recent advancements in the technologies (Gao and Chabini, 2006;Pi and Qian, 2017).…”

Section: Introductionmentioning

confidence: 99%

An optimal control approach to day-to-day congestion pricing for stochastic transportation networks

Gehlot

Honnappa

Ukkusuri

2020

Computers & Operations Research

View full text Add to dashboard Cite

Congestion pricing has become an effective instrument for traffic demand management on road networks. This paper proposes an optimal control approach for congestion pricing for day-today timescale that incorporates demand uncertainty and elasticity. Travelers make the decision to travel or not based on the experienced system travel time in the previous day and traffic managers take tolling decisions in order to minimize the average system travel time over a long time horizon. We formulate the problem as a Markov decision process (MDP) and analyze the problem to see if it satisfies conditions for conducting a satisfactory solution analysis. Such an analysis of MDPs is often dependent on the type of state space as well as on the boundedness of travel time functions. We do not constrain the travel time functions to be bounded and present an analysis centered around weighted sup-norm contractions that also holds for unbounded travel time functions. We find that the formulated MDP satisfies a set of assumptions to ensure Bellman's optimality condition. Through this result, the existence of the optimal average cost of the MDP is shown. A method based on approximate dynamic programming is proposed to resolve the implementation and computational issues of solving the control problem. Numerical results suggest that the proposed method efficiently solves the problem and produces accurate solutions.

show abstract

A stochastic optimal control approach for real-time traffic routing considering demand uncertainties and travelers’ choice heterogeneity

Cited by 19 publications

References 32 publications

Learning to Recommend Signal Plans under Incidents with Real-Time Traffic Prediction

Learning to Recommend Signal Plans under Incidents with Real-Time Traffic Prediction

Measuring and reducing the disequilibrium levels of dynamic networks with ride-sourcing vehicle data

An optimal control approach to day-to-day congestion pricing for stochastic transportation networks

Contact Info

Product

Resources

About