arXiv subjects

Jilin Song

Publications and source records attributed to Jilin Song.

5 recordsLinked to original sources

arXiv2026

Approximate Dynamic Programming for Real-time Assignment of Extraboard Transit Operators

This study investigates real-time assignment decisions for extraboard transit operators, who are responsible for covering open work due to unexpected events such as driver absenteeism. Efficient usage of extraboard operators is critical as open work negatively affects service reliability. The problem is formulated as a Markov decision process, designed to capture its stochastic and sequential nature. Due to the problem's very large state space, an approximate policy is proposed in the form of an integer program, which maps a system state to assignment decisions such that the sum of immediate and expected future rewards is maximized. As part of off-line training, future value functions for individual operators are computed using a backward dynamic program. Then, the overestimation in the aggregate value obtained by summing individual values is corrected to account for the interaction among operators. Case studies are conducted based on the operations at a real-world transit agency. Key performance metrics including uncovered open work and extraboard utilization rates are examined for varying absenteeism rates and extraboard roster sizes. The approximate policy is shown to outperform benchmark decision rules mirroring real-world assignment strategies. Further numerical experiments are conducted to analyze different operational policies: (1) inclusion of overtime drivers in the reserve operator roster; (2) reward weights for work tasks that consider passenger wait time saved. Observations from these computational analyses provide actionable insights into extraboard sizing, overtime usage, and real-time dispatch practices at transit agencies.

math.OC↗

arXiv2024

Airport Delay Prediction with Temporal Fusion Transformers

Since flight delay hurts passengers, airlines, and airports, its prediction becomes crucial for the decision-making of all stakeholders in the aviation industry and thus has been attempted by various previous research. However, previous delay predictions are often categorical and at a highly aggregated level. To improve that, this study proposes to apply the novel Temporal Fusion Transformer model and predict numerical airport arrival delays at quarter hour level for U.S. top 30 airports. Inputs to our model include airport demand and capacity forecasts, historic airport operation efficiency information, airport wind and visibility conditions, as well as enroute weather and traffic conditions. The results show that our model achieves satisfactory performance measured by small prediction errors on the test set. In addition, the interpretability analysis of the model outputs identifies the important input factors for delay prediction.

cs.LG↗

arXiv2024

Deep Reinforcement Learning for Real-Time Ground Delay Program Revision and Corresponding Flight Delay Assignments

This paper explores the optimization of Ground Delay Programs (GDP), a prevalent Traffic Management Initiative used in Air Traffic Management (ATM) to reconcile capacity and demand discrepancies at airports. Employing Reinforcement Learning (RL) to manage the inherent uncertainties in the national airspace system-such as weather variability, fluctuating flight demands, and airport arrival rates-we developed two RL models: Behavioral Cloning (BC) and Conservative Q-Learning (CQL). These models are designed to enhance GDP efficiency by utilizing a sophisticated reward function that integrates ground and airborne delays and terminal area congestion. We constructed a simulated single-airport environment, SAGDP_ENV, which incorporates real operational data along with predicted uncertainties to facilitate realistic decision-making scenarios. Utilizing the whole year 2019 data from Newark Liberty International Airport (EWR), our models aimed to preemptively set airport program rates. Despite thorough modeling and simulation, initial outcomes indicated that the models struggled to learn effectively, attributed potentially to oversimplified environmental assumptions. This paper discusses the challenges encountered, evaluates the models' performance against actual operational data, and outlines future directions to refine RL applications in ATM.

cs.LG↗

arXiv2024

Flight Path Optimization with Optimal Control Method

This paper is based on a crucial issue in the aviation world: how to optimize the trajectory and controls given to the aircraft in order to optimize flight time and fuel consumption. This study aims to provide elements of a response to this problem and to define, under certain simplifying assumptions, an optimal response, using Constrained Finite Time Optimal Control(CFTOC). The first step is to define the dynamic model of the aircraft in accordance with the controllable inputs and wind disturbances. Then we will identify a precise objective in terms of optimization and implement an optimization program to solve it under the circumstances of simulated real flight situation. Finally, the optimization result is validated and discussed by different scenarios.

math.OC↗

arXiv2024

Smart Navigation System for Parking Assignment at Large Events: Incorporating Heterogeneous Driver Characteristics

Parking challenges escalate significantly during large events such as concerts or sports games, yet few studies address dynamic parking lot assignments for such occasions. This paper introduces a smart navigation system designed to optimize parking assignments swiftly during large events, utilizing a mixed search algorithm that accounts for the heterogeneous characteristics of drivers. We conducted simulations in the Berkeley city area during the "Big Game" to validate our system and demonstrate the benefits of our innovative parking assignment approach.

cs.RO↗