The gate assignment problem (GAP) is a fundamental operational challenge for airport management, given that its solution has direct impacts on airport efficiencies, passenger experience, and resource utilization. Current studies are predominantly focused on devising problem-specific heuristics, using exact algorithms / commercial solvers, which frequently suffer from getting stuck in local optima and excessive computational times, respectively. In this paper, we propose a neural combinatorial optimization (NCO) framework to solve the GAP, which can reach a balance between computational efficiency and optimality. We embed flight and gate information into hidden states, using a multilayer transformer-based encoder-decoder policy network to assign gates in an auto-regressive manner. We train multiple NCO models using reinforcement learning and compare them to traditional exact and heuristic methods. Results show that our framework, NCOGAP, can provide practical real-time decision support with superior optimality, which we believe can greatly reduce the operator’s workload, while balancing global gate assignment operational efficiency and robustness, subject to realistic constraints and rules.

Recent prevailing works on graph machine learning typically follow a similar methodology that involves designing advanced variants of graph neural networks (GNNs) to maintain the superior performance of GNNs on different graphs. In this paper, we aim to streamline the GNN design process and leverage the advantages of Large Language Models (LLMs) to improve the performance of GNNs on downstream tasks. We formulate a new paradigm, coined "LLMs-as-Consultants", which integrates LLMs with GNNs in an interactive manner. A framework named LOGIN (LLM cOnsulted GNN traINing) is instantiated, empowering the interactive utilization of LLMs within the GNN training process. First, we attentively craft concise prompts for spotted nodes, carrying comprehensive semantic and topological information, and serving as input to LLMs. Second, we refine GNNs by devising a complementary coping mechanism that utilizes the responses from LLMs, depending on their correctness. We empirically evaluate the effectiveness of Lalebox1 [0.8]O Galebox1 [0.8]IN on node classification tasks across both homophilic and heterophilic graphs. The results illustrate that even basic GNN architectures, when employed within the proposed LLMs-as-Consultants paradigm, can achieve comparable performance to advanced GNNs with intricate designs. Our code is available at https://github.com/QiaoYRan/LOGIN.

Flight delay is a fundamental problem present in the global aviation system. Delay-inducing disruptions are caused by various reasons, including increased global connectivity / dependency, weather phenomena, and limited infrastructure resources. Given the excessive amount of time and money lost due to delays in aircraft operations, the prediction of flight delays has become an active research topic in recent years. Particularly, the existence of code repositories for standardized machine learning applications as well as the available data on this subject, has led to an increasing number of papers, usually comparing the performance of different new and existing delay prediction techniques. In this study, we review the contributions of papers to delay prediction in recent years. Based on a six-step comparison framework, covering many aspects, starting with data collection / processing, including, e.g., model and feature selection, and ending with evaluation, integration of various technologies, and reproducibility considerations, we find that although the current studies have put forward effective concerns, there are still some challenges in the field of delay prediction. We elaborate on how to overcome this stage by discussing a set of research directions, which hopefully help other researchers to perform better future studies and help to reduce the impact of delays on air transportation systems.

There exists a tremendous number of research surveys on various aspects of UAV logistics, mobility and monitoring tasks in the literature. These surveys have been published in distinct venues, often having a significant overlap in goals and key findings. In this study, we provide a meta review across nearly 100 extant UAV surveys and overview papers, extract their key messages, and investigate the extent of being complementary. We develop the AERIAL framework, which aggregates the major challenges on the way to a successful application of UAVs for logistics, mobility, and monitoring. We believe that AERIAL framework and meta review contribute towards a clearer understanding of the scientific UAV landscape challenges and the identification of potential directions for future research studies.

The choice of hub locations is fundamental in various transportation and communication systems, enabling transshipment, sorting, and consolidation functions. Recent research has highlighted the benefits of stratification when searching for hub facilities, taking into account the fact that the demand at each site belongs to different strata, e.g., depending on the type or quality of service the passengers or cargo require. In this study, we propose various instances of the stratified single allocation hub location problems, ranging from simple p-hub median / stratified hub location to extensions concerning direct links as well as node capacities. Moreover, four generalized problems, considering service level requirements, capacitated direct links, multi-modal hub networks, and incomplete network structures are derived from the basic problems. Given that the stratified models become intractable to solve for even medium-sized instances, we also develop a novel neighborhood-search based algorithm. Extensive experiments are performed on representative data sets, confirming the effectiveness and efficiency of our solution techniques. The results also further highlight the significance of stratification in hub-and-spoke networks. Our work complements recent studies on stratified p-hub center / maximal covering problems.

The data-hungry problem, characterized by insufficiency and low-quality of data, poses obstacles for deep learning models. Transfer learning has been a feasible way to transfer knowledge from high-quality external data of source domains to limited data of target domains, which follows a domain-level knowledge transfer to learn a shared posterior distribution. However, they are usually built on strong assumptions, e.g., the domain invariant posterior distribution, which is usually unsatisfied and may introduce noises, resulting in poor generalization ability on target domains. Inspired by Graph Neural Networks (GNNs) that aggregate information from neighboring nodes, we redefine the paradigm as learning a knowledge-enhanced posterior distribution for target domains, namely Knowledge Bridge Learning (KBL). KBL first learns the scope of knowledge transfer by constructing a Bridged-Graph that connects knowledgeable samples to each target sample and then performs sample-wise knowledge transfer via GNNs.KBL is free from strong assumptions and is robust to noises in the source data. Guided by KBL, we propose the Bridged-GNN including an Adaptive Knowledge Retrieval module to build Bridged-Graph and a Graph Knowledge Transfer module. Comprehensive experiments on both un-relational and relational data-hungry scenarios demonstrate the significant improvements of Bridged-GNN compared with SOTA methods

Graphs consisting of vocal nodes ("the vocal minority") and silent nodes ("the silent majority"), namely VS-Graph, are ubiquitous in the real world. The vocal nodes tend to have abundant features and labels. In contrast, silent nodes only have incomplete features and rare labels, e.g., the description and political tendency of politicians (vocal) are abundant while not for ordinary civilians (silent) on the twitter’s social network. Predicting the silent majority remains a crucial yet challenging problem. However, most existing Graph Neural Networks (GNNs) assume that all nodes belong to the same domain, without considering the missing features and distribution-shift between domains, leading to poor ability to deal with VS-Graph. To combat the above challenges, we propose Knowledge Transferable Graph Neural Network (KTGNN), which models distribution-shifts during message passing and learns representation by transferring knowledge from vocal nodes to silent nodes. Specifically, we design the domain-adapted "feature completion and message passing mechanism" for node representation learning while preserving domain difference. And a knowledge transferable classifier based on KL-divergence is followed. Comprehensive experiments on real-world scenarios (i.e., company financial risk assessment and political elections) demonstrate the superior performance of our method. Our source code has been open-sourced1.

Targeting at improving door-to-door accessibility, efficiency and reducing environmental impact, recent decades have witnessed vigorous development of multimodal transport. Coupled through passenger transfer, the failure of one mode, however, is likely to cause cascading disruptions to the complete system. Therefore, to maintain safe and efficient operations, the integrated recovery of the multimodal transportation system becomes an important, timely topic, given increasingly complex schedule interactions. This study proposes a mathematical model to handle the integrated recovery problem of multimodal transportation network by minimizing the passenger travel cost and mode recovery cost under failures of multiple critical infrastructures. For a case study on air/HSR interaction in China, our model incorporates decisions in terms of aircraft recovery, railway timetable rescheduling, and passenger routing with realistic constraints (e.g. max travel time, max transfer time). A column-and-row generation based heuristic solution algorithm is developed in order to solve the complex model with given computational budgets. Results reveal that real-world test disruption scenarios can be solved within 20 minutes with a small relaxation gap. A significant amount of passenger traveling time can be reduced in contrast to a sequential recovery approach with independent reaction decisions for each transport mode. This work contributes towards the development of support tools for integrated disruption management of multimodal transportation.

Company financial risk is ubiquitous and early risk assessment for listed companies can avoid considerable losses. Traditional methods mainly focus on the financial statements of companies and lack the complex relationships among them. However, the financial statements are often biased and lagged, making it difficult to identify risks accurately and timely. To address the challenges, we redefine the problem as company financial risk assessment on tribe-style graph by taking each listed company and its shareholders as a tribe and leveraging financial news to build inter-tribe connections. Such tribe-style graphs present different patterns to distinguish risky companies from normal ones. However, most nodes in the tribe-style graph lack attributes, making it difficult to directly adopt existing graph learning methods (e.g., Graph Neural Networks(GNNs)). In this paper, we propose a novel Hierarchical Graph Neural Network (TH-GNN) for Tribe-style graphs via two levels, with the first level to encode the structure pattern of the tribes with contrastive learning, and the second level to diffuse information based on the inter-tribe relations, achieving effective and efficient risk assessment. Extensive experiments on the real-world company dataset show that our method achieves significant improvements on financial risk assessment over previous competing methods. Also, the extensive ablation studies and visualization comprehensively show the effectiveness of our method.

On-demand air mobility services, often called air taxis, are on the way to revolutionize our urban/regional transportation sector by lifting transportation to the third dimension and thus possibly contribute to solving the congestion-induced transportation deadlock many metropolitan regions face today. Although existing research mainly focuses on the design of efficient vehicles and specifically battery technology, in the near future, a new question will arise: Where to locate the vertiports/landing pads for such air taxis? In this study, we propose a vertiport location selection problem. In contrast to existing studies, we allow the demand to be distributed over the whole metropolitan area, modeled as a grid, and exclude certain grid cells from becoming hubs, for example, because of safety/geographical constraints. The combination of these two contributions makes the problem intriguingly difficult to solve with standard solution techniques. We propose a novel variable neighborhood search heuristic, which is able to solve 12 × 12 grid instances within a few seconds of computation time and zero gaps in our experiments, whereas CPLEX needs up to 10 hours. We believe that our study contributes toward the scalable selection of vertiport locations for air taxis. Summary of Contribution: The increasing interest in opening the third dimension, that is, altitude, to transportation inside metropolitan regions raises new research challenges. Existing research mainly focuses on the design of efficient vehicles and control problems. In the near future, however, the actual operation of air taxis will lead to new set of operations research problems for so-called air taxi operations. Our contribution focuses on the optimization of vertiports for air taxi operations in a metropolitan region. We choose to model the problem over a grid-like demand structure, with a novel side constraint: selected grid cells are unavailable as hubs, for example, because of environmental, technical, cultural, or other reasons. This makes our model a special case in between the two traditional models: discrete location and continuous location. Our model is inherently difficult to solve for exact methods; for instance, solving a grid of 12 × 12 grid cells needs more than 10 hours with CPLEX, when modeled as a discrete location problem. We show that a straightforward application of existing neighborhood search heuristics is not suitable to solve this problem well. Therefore, we design an own variant of mixed variable neighborhood search, which consists of novel local search steps, tailored toward our grid structure. Our evaluation shows that by using our novel heuristic, almost all instances can be solved toward optimality.

Well-designed multi-modal transportation networks are crucial for our connected world. For instance, the excessive construction of railway tracks in China, at speeds up to 350 km/h, makes it necessary to consider the interaction of rail with air transportation for network design. In this study, we propose a model for an air/rail multi-modal, multiple allocation hub location problem with uncertainty on travel demands. Our model is unique in that it integrates features from the existing literature on multi-modal hub location problem (including hub-level capacities, link capacities, direct links, travel cost and time, transit costs and uncertainty), which have not been considered simultaneously, given its high computational complexity. We formulate this model with$O(n^{4})$variables and show that the implementation of a Benders decomposition algorithm is inherently hard, because of the cubic number of variables in the master problem. Furthermore, we derive an iterative network design algorithm and additional improvement strategies: MMHUBBI which resolves a restricted problem by the solver CPLEX and MMHUBBI-DIRECT which re-designs the transportation network by a heuristic. Our evaluation on real-world dataset for Chinese domestic transportation shows that MMHUBBI provides a significant speed-up on all instances, compared to using CPLEX, while obtaining near-optimal solutions. MMHUBBI-DIRECT further reduces the runtime/memory usage but provides solutions with worse quality. We believe that our study contributes towards the design of more realistic multi-modal hub location problems.

Federated learning is a collaborative machine learning framework where a global model is trained by different organizations under the privacy restrictions. Promising as it is, privacy and robustness issues emerge when an adversary attempts to infer the private information from the exchanged parameters or compromise the global model. Various protocols have been proposed to counter the security risks, however, it becomes challenging when one wants to make federated learning protocols robust against Byzantine adversaries while preserving the privacy of the individual participant. In this article, we propose a differentially private Byzantine-robust federated learning scheme (DPBFL) with high computation and communication efficiency. The proposed scheme is effective in preventing adversarial attacks launched by the Byzantine participants and achieves differential privacy through a novel aggregation protocol in the shuffle model. The theoretical analysis indicates that the proposed scheme converges to the approximate optimal solution with the learning error dependent on the differential privacy budget and the number of Byzantine participants. Experimental results on MNIST, FashionMNIST and CIFAR10 demonstrate that the proposed scheme is effective and efficient.

This study aims to clarify a few concepts when assessing the robustness of complex systems with the usage of complex network methods. Many researchers simply use the degree centrality as a proxy for node importance in a network. We show that such assumption can lead to rather overestimated robustness values, neglecting much stronger attacks.

The introduction of ADS-B, a satellite-based aircraft tracking technology, and the increasing installation of ADS-B receiver stations around the globe eases the tracking of aircraft, compared with traditional solutions using secondary radar. Given the large scale of ADS-B implementation and the high frequency of data collection, storing and managing ADS-B induced data has become increasingly difficult: The worldwide ADS-B data easily aggregates to several hundreds of terabyte per year, depending on the spatial coverage and temporal resolution. Standard data management solutions do not work well for ADS-B data, since they either require a large uncompressed index structure or cannot be queried efficiently. In this paper, we propose a novel compressed index structure for managing ADS-B data, called ADS-BI. The essential building blocks are spatio-temporal reference partitioning, reordering, and compression. On top of the partitioned, compressed representation, metadata is stored effectively, and exploited during query answering for typical ATM related task such as trajectory adherence evaluation, as well as complexity and safety metrics assessment by only accessing parts of the compressed data as necessary. Our novel index structure is evaluated on worldwide ADS-B data for a week in November 2016. For comparison, we implemented ten standard compression/indexing methods. The experiments reveal that none of these traditional methods can target the sweet spot between a small storage and efficient query answering. Our novel technique provides fast query answering at smallest storage costs. This paper contributes toward efficient handling of the increasing amount of traffic data in air traffic management, and eventually, toward more efficient and safer air transportation.

Understanding and improving global mobility has gained increased interest during the last decades. However, studies on the railway network are spatially limited so far, mostly investigating the domestic network of a country. Data availability is a major limiting factor for the analysis of these networks. Despite the increased open data movement, network operators are often reluctant to publish their infrastructure and passenger data. Existing large-scale studies usually make use of hand-collected data, for instance, based on historical cartographies. In this paper, we develop and implement a methodology to extract the worldwide railway skeleton network from the open data repository OpenStreetMap, where nodes are stations/waypoints and links are weights with information such as spatial distance, gauge, and maximum speed. We describe how we solved several data cleansing and scalability issues and developed network simplification techniques, in order to obtain an adequate representation of the network. We show that the network breaks down into few large and many small components. Furthermore, we show that this public data set can be used for efficient minimum travel time estimation between stations or cities. This paper leads to the development of a new research data set and contributes toward the ability of analyzing global mobility patterns, particularly regarding multimodality and cross-country transportation.

Transit agencies electronically publish transit schedules and route data to improve customer experience or as part of the open-data initiative. The release of such data poses several data management challenges because schedules for a single city can easily exceed storage requirements of several hundreds of megabytes. One way to deal with these challenges is data compression. The encoding of public transit schedules often follows the General Transit Feed Specification (GTFS), which cannot be compressed well out of the box. We propose GTFSCompress, which is an algorithm for compression of GTFS data, based on referential compression queues, which compress a column stream depending on previously seen items in this stream and other streams. Our evaluation on ten real-world data sets shows that GTFS feeds can be compressed by a factor of 100 and more. This is up to one order of magnitude better than using the best standard compressors.

Air traffic management (ATM) is facing a tremendous increase in the amount of available flight data, particularly four-dimensional (4D) trajectories. Computational requirements for analysis and storage of such bulk of data are steeply increasing. Compression is one key technology to address this challenge. In this paper we propose two techniques for compressing air traffic 4D trajectories. Our first technique analyzes a set of samples and computes a prediction for the most likely picked successor coordinate by a random walker. The second technique, i.e., referential compression, compresses a 4D trajectory as a collection of subtrajectory pointers into a reference trajectory. We evaluate our algorithms on trajectory data from the Demand Data Repository provided by EUROCONTROL. We show that a combination of our referential and statistical compression techniques compresses 4D trajectories of all air traffic over Europe in the year 2013 from 60 GB down to 0.78 GB, achieving a compression ratio of more than 75 : 1. The compression ratio for our techniques increases with the number of to-be-compressed flights, whereas standard compression techniques achieve a fixed compressed ratio for any number of flights. Our work contributes toward efficient handling of the increasing amount of traffic data in ATM.

Identification of genes and pathways involving in diseases and physiological conditions is a major task in systems biology. In this study, we develop a new non-parameter Ising model to integrate protein-protein interaction network and microarray data for identifying differentially expressed (DE) genes. We also propose a simulated annealing algorithm to find the optimal configuration of the Ising model. We test the Ising model to two breast cancer microarray data sets. The results show that more cancer related differentially expressed subnetworks and genes are identified by the Ising model than by the Markov random filed (MRF) model.