Large Language Models, particularly decoder-only generative models such as GPT, are increasingly used to automate Software Engineering tasks. These models are primarily guided through natural language prompts, making prompt engineering a critical factor in system performance and behavior. Despite their growing role in SE research, prompt-related decisions are rarely documented in a systematic or transparent manner, hindering reproducibility and comparability across studies. To address this gap, we conducted a two-phase empirical study. First, we analyzed nearly 300 papers published at the top-3 SE conferences since 2022 to assess how prompt design, testing, and optimization are currently reported. Second, we surveyed 105 program committee members from these conferences to capture their expectations for prompt reporting in LLM-driven research. Based on the findings, we derived a structured guideline that distinguishes essential, desirable, and exceptional reporting elements. Our results reveal significant misalignment between current practices and reviewer expectations, particularly regarding version disclosure, prompt justification, and threats to validity. We present our guideline as a step toward improving transparency, reproducibility, and methodological rigor in LLM-based SE research.
Alexandra González, Xavier Franch, Silverio Martínez-Fernández
Integrating Artificial Intelligence into Software Engineering (SE) requires having a curated collection of models suited to SE tasks. With millions of models hosted on Hugging Face (HF) and new ones continuously being created, it is infeasible to identify SE models without a dedicated catalogue. To address this gap, we present SEMODS: an SE-focused dataset of 3,427 models extracted from HF, combining automated collection with rigorous validation through manual annotation and large language model assistance. Our dataset links models to SE tasks and activities from the software development lifecycle, offering a standardized representation of their evaluation results, and supporting multiple applications such as data analysis, model discovery, benchmarking, and model adaptation.
Mauro Marcelino, Marcos Alves, Bianca Trinkenreich
et al.
[Context] An evidence briefing is a concise and objective transfer medium that can present the main findings of a study to software engineers in the industry. Although practitioners and researchers have deemed Evidence Briefings useful, their production requires manual labor, which may be a significant challenge to their broad adoption. [Goal] The goal of this registered report is to describe an experimental protocol for evaluating LLM-generated evidence briefings for secondary studies in terms of content fidelity, ease of understanding, and usefulness, as perceived by researchers and practitioners, compared to human-made briefings. [Method] We developed an RAG-based LLM tool to generate evidence briefings. We used the tool to automatically generate two evidence briefings that had been manually generated in previous research efforts. We designed a controlled experiment to evaluate how the LLM-generated briefings compare to the human-made ones regarding perceived content fidelity, ease of understanding, and usefulness. [Results] To be reported after the experimental trials. [Conclusion] Depending on the experiment results.
Systems engineering (SE) is evolving with the availability of generative artificial intelligence (AI) and the demand for a systems-of-systems perspective, formalized under the purview of mission engineering (ME) in the US Department of Defense. Formulating ME problems is challenging because they are open-ended exercises that involve translation of ill-defined problems into well-defined ones that are amenable for engineering development. It remains to be seen to which extent AI could assist problem formulation objectives. To that end, this paper explores the quality and consistency of multi-purpose Large Language Models (LLM) in supporting ME problem formulation tasks, specifically focusing on stakeholder identification. We identify a relevant reference problem, a NASA space mission design challenge, and document ChatGPT-3.5's ability to perform stakeholder identification tasks. We execute multiple parallel attempts and qualitatively evaluate LLM outputs, focusing on both their quality and variability. Our findings portray a nuanced picture. We find that the LLM performs well in identifying human-focused stakeholders but poorly in recognizing external systems and environmental factors, despite explicit efforts to account for these. Additionally, LLMs struggle with preserving the desired level of abstraction and exhibit a tendency to produce solution specific outputs that are inappropriate for problem formulation. More importantly, we document great variability among parallel threads, highlighting that LLM outputs should be used with caution, ideally by adopting a stochastic view of their abilities. Overall, our findings suggest that, while ChatGPT could reduce some expert workload, its lack of consistency and domain understanding may limit its reliability for problem formulation tasks.
Khandakar Ashrafi Akbar, Md Nahiyan Uddin, Latifur Khan
et al.
As connected and automated transportation systems evolve, there is a growing need for federal and state authorities to revise existing laws and develop new statutes to address emerging cybersecurity and data privacy challenges. This study introduces a Retrieval-Augmented Generation (RAG) based Large Language Model (LLM) framework designed to support policymakers by extracting relevant legal content and generating accurate, inquiry-specific responses. The framework focuses on reducing hallucinations in LLMs by using a curated set of domain-specific questions to guide response generation. By incorporating retrieval mechanisms, the system enhances the factual grounding and specificity of its outputs. Our analysis shows that the proposed RAG-based LLM outperforms leading commercial LLMs across four evaluation metrics: AlignScore, ParaScore, BERTScore, and ROUGE, demonstrating its effectiveness in producing reliable and context-aware legal insights. This approach offers a scalable, AI-driven method for legislative analysis, supporting efforts to update legal frameworks in line with advancements in transportation technologies.
Ahmed Danladi Abdullahi, Erfan Bahrami, Tooska Dargahi
et al.
The advancement of 6G technology has the potential to revolutionize the transportation sector and significantly improve how we travel. 6G-enabled Intelligent Transportation Systems (ITS) promise to offer high-speed, low-latency communication and advanced data analytics capabilities, supporting the development of safer, more efficient, and more sustainable transportation solutions. However, various security and privacy challenges were identified in the literature that must be addressed to enable the safe and secure deployment of 6G-ITS and ensure people's trust in using these technologies. This paper reviews the opportunities and challenges of 6G-ITS, particularly focusing on trust, security, and privacy, with special attention to quantum technologies that both enhance security through quantum key distribution and introduce new vulnerabilities. It discusses the potential benefits of 6G technology in the transportation sector, including improved communication, device interoperability support, data analytic capabilities, and increased automation for different components, such as transportation management and communication systems. A taxonomy of different attack models in 6G-ITS is proposed, and a comparison of the security threats in 5G-ITS and 6G-ITS is provided, along with potential mitigating solutions. This research highlights the urgent need for a comprehensive, multi-layered security framework spanning physical infrastructure protection, network protocol security, data management safeguards, application security measures, and trust management systems to effectively mitigate emerging security and privacy risks and ensure the integrity and resilience of future transportation ecosystems.
Improving public transport connectivity is crucial for decarbonisation and economic growth. Current transport planning approaches to addressing connectivity problems rely on trial-and-error approaches to identify problems and generate options, limited by planners' incomplete knowledge and the overwhelming volume of available travel data.We introduce a machine-assisted approach to identify opportunities for connectivity enhancements from origin-destination data and generate prioritised intervention options. Using an origin-destination matrix for Greater London covering approximately 1200 activity centres, our method applies trajectory clustering to identify potential high-demand corridors with poor public transport quality.Our prototype automatically generates multiple public transport scheme options (local bus, express bus, metro) within these corridors along with approximate operating costs. These options are batch-tested using accelerated assignment modelling that optimises mode choice, frequency, and route generation, and the results are given ordered according to benefit-cost ratios.This approach is now being used to supplement human planners’ knowledge in the development of new express bus services in London.
Transportation and communications, Transportation engineering
This paper presents a comprehensive and up-to-date review of energy dissipative devices for controlling vortex-induced vibrations (VIV) in bridge decks, covering passive, active, and semi-active control systems. It synthesizes key findings and innovative control strategies from the VIV control research community, critically evaluating the advantages and limitations of each device type and highlighting essential research gaps for future exploration. A systematic literature review identifies and summarizes twelve documented real-bridge applications of these devices: ten involving tuned mass dampers (TMDs), one with viscous dampers, and one employing semi-active TMDs. While TMDs have demonstrated effectiveness, ongoing research focuses on optimizing their parameters under nonlinear vortex-induced forces, closely spaced modal frequencies, and multi-objective criteria. Emerging research trends emphasize the development of advanced passive devices for ultra-low-frequency VIV control, lightweight and compact designs, and multimodal vibration mitigation. Furthermore, active and semi-active control devices show considerable promise due to their superior adaptability and enhanced performance over passive systems, indicating a promising direction for future VIV mitigation strategies in long-span bridges.
[Objective] Tunnel deformation is related to the health of the tunnel structure, and accurate monitoring of tunnel deformation is very important for tunnel safety. Although traditional manual measurement methods are relatively accurate, the time-consuming and labor-intensive defects making it difficult to meet the efficient operation and maintenance requirements of large-scale tunnels; while automatic monitoring methods are mainly based on imported fully automatic total stations, with limited measurement range and high equipment costs. It is urgent to develop new efficient and low-cost measurement technologies. [Method] Based on machine vision measurement, large-scale multi-point deformation monitoring of subway tunnel structures can be achieved in a single-camera mode, which can take into account both monitoring frequency and accuracy, featuring the advantages of being simple and fast. First, after selecting the camera parameters, the magnification at each monitoring point is calibrated; at the same time, the grayscale centroid method is used to quickly calculate the center coordinates of the light spot and its change during the monitoring process; finally, the magnification is used to realize the conversion of pixel coordinate changes to actual physical coordinates, thereby obtaining the actual displacement deformation of each monitoring point. [Result & Conclusion] Based on machine vision methods, a 28-day displacement monitoring experiment is conducted on a 120m subway tunnel section. The experimental results show that both the horizontal and vertical displacements of the target tunnel section are within 1.5mm, and both exhibit an overall trend of periodic fluctuations. This verifies that machine vision monitoring of subway tunnels can achieve long-term continuous and accurate monitoring of millimeter-level subway tunnel deformation.
This study asserts that paired aircraft can withstand specific wake turbulence levels and explores the longitudinal collision risk in closely spaced parallel runway approaches. The goal is to enhance the safety margin of the paired approach and allow for more flexible implementation. Based on QAR data, a theoretical spacing model for paired aircraft and a probability distribution of acceleration error are established to facilitate the analysis of the actual spacing of paired aircraft. Wake turbulence attenuation is modelled using large eddy simulation, creating a vortex attenuation model. Drawing inspiration from the Hallock-Burnham vortex model, new models for induced velocity and vortex core motion are proposed. The study assumes that trailing aircraft can handle certain wake intensities, leading to a new model for calculating wake turbulence safety intervals, limiting the trailing aircraft’s maximum roll angle to its critical limit. Using probability theory, a model for longitudinal collision risk is formulated, combining wake turbulence safety separation and the actual separation of paired aircraft. The study also examines various factors influencing longitudinal collision risk, emphasising the significant impact of crosswind conditions. It concludes that a stronger crosswind component reduces the wake turbulence safety separation, thereby increasing the risk of longitudinal collisions, particularly during the final stage of the approach. Notably, collision risk is directly proportional to the crosswind component and initial longitudinal separation, but inversely proportional to runway spacing.
Szabolcs Duleba, Bálint Farkas, Sarbast Moslem
et al.
Rail freight policy should reflect both the current and foreseen situations of the certain national rail freight market. The situation analysis is often based on expert evaluations, in many cases with the participation of the market players themselves, and their synthesized group opinion is the basis of policy-making. However, the creation of opinion synthesis has not considered the market power of the players so far, and a clear research gap exists on how to properly addressing weights to the respondent experts in the group of evaluators to gain a realistic image on the present and future of the rail freight market. The objective of this paper is to identify problems, risks, and development potential in an EU national rail freight market by an expert survey based new methodology (Sectoral Group Analytic Hierarchy Process, SGAHP) that assigns different decision-maker weights in the respondent group based on the different market power of the players to gain a clear and overall image on the examined market. As a case study, a survey has been conducted in Hungary involving the significant representatives of the national rail freight. Results show that the recruitment and training of human resource is a common problem, while supporting single wagon traffic is meaningless for all players. However, big companies prioritise the reduction of locomotive maintenance time, while small ones strive to have own maintenance facilities. As an implication, the proposed model might help state decision-makers in customizing financial or other support to efficiently increase the competitiveness of the sector, as well as the rail companies to better adopt to the situation. The proposed new model has been proven successful not only from the aspect of robustness and sensitivity, but also of recommending practical modifications in rail freight to transport planners both on national and EU levels.
Abstract Recently, solar-driven gas-liquid interfacial evaporation has mainly concentrated on the improvement of sunlight absorption efficiency while neglecting the regulation of water transportation. The blind pursuit of superwetting material design has ignored using excessive water transport instead to add unnecessary heat loss. Furthermore, the material biosafety also should be focused which can avoid damage to water-source biology and humans during application. Here, we successfully designed a biosafety cuttlefish juice (CJ)-based solar absorber via rotary filling SiO2 nanoparticles (NPs) to optimize the water-transport regulation of SGIE systems, the tailored absorber can block excessive useless water, and effectively reduce heat losses. Hence the absorber has specific and significant advantages: higher evaporation efficiency and fast evaporation response to the entire device, excellent biocompatibility and low-toxicity, IC50 more than 100 mg/ml is observed in the cell viability. In addition, high mechanical properties (pressure≈100 N), multi-size preparation (d ~ 4.25 cm–10.25 cm or bigger), and long-time self-floating (more than 720 h) yield advantages together to achieve an excellent performance evaporator. Therefore, this work provides a new avenue to the design of high performance SGIE by enchanting edge zone effect, controlling water transportation and fabricating with sustainable marine resource.
Marco Autili, Martina De Sanctis, Paola Inverardi
et al.
As testified by new regulations like the European AI Act, worries about the human and societal impact of (autonomous) software technologies are becoming of public concern. Human, societal, and environmental values, alongside traditional software quality, are increasingly recognized as essential for sustainability and long-term well-being. Traditionally, systems are engineered taking into account business goals and technology drivers. Considering the growing awareness in the community, in this paper, we argue that engineering of systems should also consider human, societal, and environmental drivers. Then, we identify the macro and technological challenges by focusing on humans and their role while co-existing with digital systems. The first challenge considers humans in a proactive role when interacting with digital systems, i.e., taking initiative in making things happen instead of reacting to events. The second concerns humans having a reactive role in interacting with digital systems, i.e., humans interacting with digital systems as a reaction to events. The third challenge focuses on humans with a passive role, i.e., they experience, enjoy or even suffer the decisions and/or actions of digital systems. The fourth challenge concerns the duality of trust and trustworthiness, with humans playing any role. Building on the new human, societal, and environmental drivers and the macro and technological challenges, we identify a research roadmap of digital systems for humanity. The research roadmap is concretized in a number of research directions organized into four groups: development process, requirements engineering, software architecture and design, and verification and validation.
Ubiquitous mobile devices have catalyzed the development of vehicle crowd sensing (VCS). In particular, vehicle sensing systems show great potential in the flexible acquisition of spatio-temporal urban data through built-in sensors under diverse sensing scenarios. However, vehicle systems often exhibit biased coverage due to the heterogeneous nature of trip requests and routes. To achieve a high sensing coverage, a critical challenge lies in optimally relocating vehicles to minimize the divergence between vehicle distributions and target sensing distributions. Conventional approaches typically employ a two-stage predict-then-optimize (PTO) process: first predicting real-time vehicle distributions and subsequently generating an optimal relocation strategy based on the predictions. However, this approach can lead to suboptimal decision-making due to the propagation of errors from upstream prediction. To this end, we develop an end-to-end Smart Predict-then-Optimize (SPO) framework by integrating optimization into prediction within the deep learning architecture, and the entire framework is trained by minimizing the task-specific matching divergence rather than the upstream prediction error. Methodologically, we formulate the vehicle relocation problem by quadratic programming (QP) and incorporate a novel unrolling approach based on the Alternating Direction Method of Multipliers (ADMM) within the SPO framework to compute gradients of the QP layer, facilitating backpropagation and gradient-based optimization for end-to-end learning. The effectiveness of the proposed framework is validated by real-world taxi datasets in Hong Kong. Utilizing the alternating differentiation method, the general SPO framework presents a novel concept of addressing decision-making problems with uncertainty, demonstrating significant potential for advancing applications in intelligent transportation systems.
Abstract This study addresses the challenge of quantitatively assessing ride comfort in tram travel in Growing Urban Environments, where multiple influencing factors complicate developing a unified evaluation index system. A comprehensive evaluation framework based on cloud theory is proposed to overcome this challenge. The approach involves defining five‐level comfort evaluation grades to capture passengers' experiences and perceptions accurately. The Criteria Importance through Inter‐Criteria Correlation (CRITIC) method is employed to ensure objectivity to establish objective weights for evaluation indices. Subsequently, a cloud model algorithm is utilized to generate evaluation benchmark and actual result clouds, providing intuitive representations of the evaluation outcomes. The efficacy and rationality of the methodology is illustrated through a case study focusing on Suzhou Tram Line 2. This research contributes valuable insights for enhancing public transportation experiences in new urban settings by offering a systematic and objective approach to assessing tram ride comfort.
Urban populations, especially women, are vunerable to exposure to airborne pollution, particularly inhalable particulates (PM10). Thus, more accurate measurement of PM10 levels and evaluating their health effects is critical for guiding policy to improve human health. Previous studies obtained personal PM10 with time-weighted average by air filter-based sampling (AFS), which ignores individual differences and behavioral patterns. Here, we used nasal filters instead of AFS to obtain actual inhaled PM10 under short-term exposure for urban dwelling women during a severe haze event in Beijing in 2016. The levels of six heavy metals such as As, Cd, Ni, Cr, Pb, and Co in PM10 were investigated, and carcinogenic and non-carcinogenic risks evaluated based on an adjusted US EPA health risk assessment model. The health endpoints for urban dwelling women were further assessed through an exposure-reponse model. We found that the hourly inhaled dose of PM10 obtained through the nasal filter was about 2.5–17.6 times that obtained by AFS, which also resulted in 4.41–11.30 times more morbidity than estimated by AFS (p < 0.05). Proximity to traffic emissions resulted in greater exposure to particulate matter (>18.8 μg/kg·h) and heavy metals (>2.2 ng/kg·h), and these populations are therefore at greatest risk of developing non-cancer (HI = 4.16) and cancer (Rt = 7.8 × 10−3) related morbities.
Wiparat Siripong, Philip F. Wolf, Theodora Puspowangi Kusumoputri
et al.
BackgroundInterests in renewable fuels have exploded in recent years as the serious effects of global climate change become apparent. Microbial production of high-energy fuels by economically efficient bioprocesses has emerged as an attractive alternative to the traditional production of transportation fuels. Here, we engineered Pichia pastoris, an industrial workhorse in heterologous enzyme production, to produce the biofuel isobutanol from two renewable carbon sources, glucose and glycerol. Our strategy exploited the yeast’s amino acid biosynthetic pathway and diverted the amino acid intermediates to the 2-keto acid degradation pathway for higher alcohol production. To further demonstrate the versatility of our yeast platform, we incorporated a broad-substrate-range alcohol-O-acyltransferase to generate a variety of volatile esters, including isobutyl acetate ester and isopentyl acetate ester.ResultsThe engineered strain overexpressing the keto-acid degradation pathway was able to produce 284 mg/L of isobutanol when supplemented with 2-ketoisovalerate. To improve the production of isobutanol and eliminate the need to supplement the production media with the expensive 2-ketoisovalerate intermediate, we overexpressed a portion of the amino acid l-valine biosynthetic pathway in the engineered strain. While heterologous expression of the pathway genes from the yeast Saccharomyces cerevisiae did not lead to improvement in isobutanol production in the engineered P. pastoris, overexpression of the endogenous l-valine biosynthetic pathway genes led to a strain that is able to produce 0.89 g/L of isobutanol. Fine-tuning the expression of bottleneck enzymes by employing an episomal plasmid-based expression system further improved the production titer of isobutanol to 2.22 g/L, a 43-fold improvement from the levels observed in the original strain. Finally, heterologous expression of a broad-substrate-range alcohol-O-acyltransferase led to the production of isobutyl acetate ester and isopentyl acetate ester at 51 and 24 mg/L, respectively.ConclusionsIn this study, we engineered high-level production of the biofuel isobutanol and the corresponding acetate ester by P. pastoris from readily available carbon sources. We envision that our work will provide an economic route to this important class of compounds and establish P. pastoris as a versatile production platform for fuels and chemicals.