We present Vertiport-informed Surrogate-Based Optimization with Machine Learning Surrogates (VinS), a novel framework for solving the vertiport location problem for urban air mobility operations. The primary focus of this work is on the optimization of vertiport locations to facilitate efficient urban air transportation. Our framework helps choose not only the optimal vertiport locations but also the optimal number of vertiports. We develop a simulation model to analyze the impacts of various vertiport location configurations on the efficiency of the transportation network. To expedite the simulation process, a surrogate model is trained using machine learning algorithms, effectively reducing the computational time for evaluating a given vertiport location configuration. With the machine learning surrogate models, we apply a genetic algorithm to find the optimal set of vertiport locations. An empirical study was performed in the San Francisco Bay Area, from which we found that given the optimal set of vertiport locations, we further reduced the total travel time in the entire transportation system in the Bay Area compared with the sampled sets by 0.05% (400 h) on average. We release our framework at: https://github.com/Xuan-1998/LPSim .

Public transit systems are critical to the quality of urban life, and enhancing their efficiency is essential for building cost-effective and sustainable smart cities. Historically, researchers sought reinforcement learning (RL) applications to mitigate bus bunching issues with holding strategies. Nonetheless, these attempts often led to oversimplifications and misalignment with the goal of reducing the total time passengers spent in the system, resulting in less robust or non-optimal solutions. In this study, we introduce a novel setting where each bus, supervised by an RL agent, can appropriately form aggregated policies from three strategies (holding, skipping station, and turning around to serve the opposite direction). It’s difficult to learn them all together, due to learning complexity, we employ domain knowledge and develop a gradually expanding action space curriculum, enabling agents to learn these strategies incrementally. We incorporate Long Short-Term Memory (LSTM) in our model considering the temporal interrelation among these actions. To address the inherent uncertainties of real-world traffic systems, we impose Domain Randomization (DR) on variables such as passenger demand and bus schedules. We conduct extensive numerical experiments with the integration of synthetic and real-world data to evaluate our model. Our methodology proves effective, enhancing bus schedule reliability and reducing total passenger waiting time by over 15%, thereby improving bus operation efficiency and smoothering operations of buses that align with sustainable goals. This work highlights the potential of robust RL combined with curriculum learning for optimizing public transport in smart cities, offering a scalable solution for real-world multi-agent systems.

Parallel discrete event simulation is a fundamental simulation technology that is essential to the parallelization of event-based models including hardware and transportation systems. Parallelization is often difficult due to dynamic data-dependencies and limited computational work for hiding runtime overheads. We present Devastator, a scalable parallel discrete event simulation framework for modern C++. Devastator provides a productive API that leverages C++’s type system to eliminate boilerplate code. Devastator has been designed to specifically optimize performance on distributed many-core architectures with deepening memory hierarchies. Devastator relies on the GASNet-Ex communication runtime as well as lock-free message queues to achieve highly competitive performance. We perform strong and weak scaling studies on NERSC’s Perlmutter up to 32,698 cores and demonstrate an up to 5x speed-up over the ROSS simulator for workloads with substantial locality.

In March 1998, a survey was conducted on trains running in Switzerland in order to gauge the public interest in utilizing a new mag-lev train system operated by Swissmetro. The survey results are often used as data in order to facilitate trip mode prediction models. In order to explore the double descent hypothesis, which posits that overparameterization of models helps generalization rather than harming it, a multinomial logit model with regu- larization is compared against neural-network models that overparameterized in this study. We finally achieved 67% accuracy on the neural-network compared with the 44% accuracy of the multinomial logit model. The models are tested against a common testing model, and the accuracy performance of the models are examined. Additionally, to further investi- gate the choice of parameters commonly used in prediction models, several causal inference studies are carried out to adjust the effect of the survey location (as a proxy for existing mode choice) and travelers’ high-income status against possible confounders. Ultimately, causal inference further strengthened the variables survey location and traveler income as effective predictors in models, especially when treated for causal variables.

This paper presents a scaling methodology for benchmarking the performance of traffic simulation algorithms on multiple GPUs with different hardware specifications and the naive implementation of GPU-based traffic simulation with multi-GPU machines. The methodology aims to enable fair comparisons of runtime across different devices and assess how a program scales with the number of cores in the GPU. The study includes a comprehensive evaluation of a traffic simulation algorithm implemented on multiple graphical processing units (GPUs), comparing its performance with a single-GPU implementation. The results demonstrate that the multi-GPU implementation outperforms the single-GPU implementation in terms of sim- ulation speed, highlighting the benefits of implementing traffic simulations on multi-GPU architectures. The findings provide insights for designing more efficient and scalable traffic simulation algorithms, contributing to improved traffic management and congestion reduction on roads. The paper also discusses the use of Bulk Synchronous Parallel (BSP) modeling, normalizing runtime across different GPU hardware, and the application of Amdahl’s law to assess speedup and optimize parallelization. Future work includes optimizing inter-GPU communication and load balancing for further scalability improvements.