Multi-Agent Multimodal Large Language Model Framework for Automated Interpretation of Fuel Efficiency Analytics in Public Transportation
Source abstract
Abstract
Enhancing fuel efficiency in public transportation requires the integration of complex multimodal data into interpretable, decision-relevant insights. However, traditional analytics and visualization methods often yield fragmented outputs that demand extensive human interpretation, limiting scalability and consistency. This study presents a multi-agent framework that leverages multimodal large language models (LLMs) to automate data narration and energy insight generation. The framework coordinates three specialized agents, including a data narration agent, an LLM-as-a-judge agent, and an optional human-in-the-loop evaluator, to iteratively transform analytical artifacts into coherent, stakeholder-oriented reports. The system is validated through a real-world case study on public bus transportation in Northern Jutland, Denmark, where fuel efficiency data from 4006 trips are analyzed using Gaussian Mixture Model clustering. Comparative experiments across five state-of-the-art LLMs and three prompting paradigms identify GPT-4.1 mini with Chain-of-Thought prompting as the optimal configuration, achieving 97.3% narrative accuracy while balancing interpretability and computational cost. The findings demonstrate that multi-agent orchestration significantly enhances factual precision, coherence, and scalability in LLM-based reporting. The proposed framework establishes a replicable and domain-adaptive methodology for AI-driven narrative generation and decision support in energy informatics.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Vehicle emissions and performance · Energy, Environment, and Transportation Policies · Transportation and Mobility Innovations
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
SDU Center for Energy Informatics, The Maersk Mc-Kinney Moller Institute, Faculty of Engineering, University of Southern Denmark, DK-5230 Odense, Denmark
Institute for Technologies and Management of Digital Transformation, University of Wuppertal, D-42119 Wuppertal, Germany
Funding
the Energy Technology Development and Demonstration Pro-gramme
References (40)
Show source-provided references
1. Sahu, G., Puri, A., Rodriguez, J., Abaskohi, A., Chegini, M., Drouin, A., Taslakian, P., Zantedeschi, V., Lacoste, A., and Vazquez, D. (2025, January 24–28). Insightbench: Evaluating business analytics agents through multi-step insight generation. Proceedings of the The Thirteenth International Conference on Learning Representations, Singapore.
2. Zhang, C., Li, C., and Gao, S. (2025). MDSF: Context-Aware Multi-Dimensional Data Storytelling Framework based on Large language Model. arXiv.
3. Luo, J., Ouyang, C., Jing, Y., Fang, H., Xiao, Y., Zhang, Q., and Li, R. (2024, January 18–20). Application of LLM Techniques for Data Insights in DHP. Proceedings of the 2024 IEEE 4th International Conference on Digital Twins and Parallel Intelligence (DTPI), Wuha, China.10.1109/DTPI61353.2024.10778677 ↗
4. Islam, M.S., Laskar, M.T., Parvez, M.R., Hoque, E., and Joty, S. (2024, January 12–16). DataNarrative: Automated data-driven storytelling with visualizations and texts. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, Miami, FL, USA.10.18653/v1/2024.emnlp-main.1073 ↗
5. Infographics and public policy: Using data visualization to convey complex information10.1377/hlthaff.2015.0642 ↗
6. Wang, L., Zhang, S., Wang, Y., Lim, E.P., and Wang, Y. (2023, January 6–10). LLM4Vis: Explainable Visualization Recommendation using ChatGPT. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing: Industry Track, Singapore.10.18653/v1/2023.emnlp-industry.64 ↗
7. Chen, D., Chen, R., Zhang, S., Wang, Y., Liu, Y., Zhou, H., Zhang, Q., Wan, Y., Zhou, P., and Sun, L. (2024, January 21–27). Mllm-as-a-judge: Assessing multimodal llm-as-a-judge with vision-language benchmark. Proceedings of the Forty-first International Conference on Machine Learning, Vienna, Austria.
8. Towards a holistic framework for multimodal LLM in 3D brain CT radiology report generation10.1038/s41467-025-57426-0 ↗
9. Rahman, R., Hasan, R., Farhad, A.A., Laskar, M.T., Ashmafee, M.H., and Kamal, A.R. (2023, January 5–9). ChartSumm: A Comprehensive Benchmark for Automatic Chart Summarization of Long and Short Summaries. Proceedings of the 36th Canadian Conference on Artificial Intelligence, Montreal, QC, Canada.10.21428/594757db.0b1f96f6 ↗
10. Xie, K., and Riedl, M. (2024, January 17–22). Creating Suspenseful Stories: Iterative Planning with Large Language Models. Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics, St. Julian’s, Malta.10.18653/v1/2024.eacl-long.147 ↗
11. Hu, A., Shi, Y., Xu, H., Ye, J., Ye, Q., Yan, M., Li, C., Qian, Q., Zhang, J., and Huang, F. (November, January 28). mplug-paperowl: Scientific diagram analysis with the multimodal large language model. Proceedings of the 32nd ACM International Conference on Multimedia, Melbourne, Australia.
12. Shen, L., Li, H., Wang, Y., and Qu, H. (2024, January 13). From data to story: Towards automatic animated data video creation with llm-based multi-agent systems. Proceedings of the 2024 IEEE VIS Workshop on Data Storytelling in an Era of Generative AI (GEN4DS), St. Pete Beach, FL, USA.10.1109/GEN4DS63889.2024.00008 ↗
13. Let storytelling tell vivid stories: A multi-modal-agent-based unified storytelling framework10.1016/j.neucom.2024.129316 ↗
14. Shojaee, P., Mirzadeh, I., Alizadeh, K., Horton, M., Bengio, S., and Farajtabar, M. (2025). The illusion of thinking: Understanding the strengths and limitations of reasoning models via the lens of problem complexity. arXiv.10.70777/si.v2i6.15919 ↗
15. Satpute, A., Gießing, N., Greiner-Petter, A., Schubotz, M., Teschke, O., Aizawa, A., and Gipp, B. (2024, January 14–18). Can llms master math? investigating large language models on math stack exchange. Proceedings of the 47th International ACM SIGIR Conference on Research and Development in Information Retrieval, Washington, DC, USA.10.1145/3626772.3657945 ↗
16. Wang, Z., Zhang, H., Li, C.L., Eisenschlos, J.M., Perot, V., Wang, Z., Miculicich, L., Fujii, Y., Shang, J., and Lee, C.Y. (2024). Chain-of-table: Evolving tables in the reasoning chain for table understanding. arXiv.
17. Ma, Z., Jorgensen, B.N., and Ma, Z. (2024). A Scoping Review of Energy-Efficient Driving Behaviors and Applied State-of-the-Art AI Methods. Energies, 17.10.3390/en17020500 ↗
18. Ma, Z., Jørgensen, B.N., and Ma, Z. (2024, January 3–6). Fuel Efficiency Analysis of the Public Transportation System Based on the Gaussian Mixture Model Clustering. Proceedings of the EPIA Conference on Artificial Intelligence, Viana do Castelo, Portugal.10.1007/978-3-031-73500-4_18 ↗
19. Zhang, D., Yu, Y., Dong, J., Li, C., Su, D., Chu, C., and Yu, D. (2024). MM-LLMs: Recent Advances in MultiModal Large Language Models. Findings of the Association for Computational Linguistics: ACL 2024, Association for Computational Linguistics.10.18653/v1/2024.findings-acl.738 ↗
20. (2025, April 14). OpenAI Introducing GPT-4.1 in the API. Available online: https://openai.com/index/gpt-4-1/.
21. Anthropic (2025, May 22). Introducing Claude 4. Available online: https://www.anthropic.com/news/claude-4.
22. Google DeepMind (2025, May 07). Gemini 2.5 Pro. Available online: https://deepmind.google/models/gemini/pro/.
23. (2025, April 16). OpenAI Introducing OpenAI o3 and o4-mini. Available online: https://openai.com/index/introducing-o3-and-o4-mini/.
24. Anthropic (2024, October 22). Introducing Computer Use, a New Claude 3.5 Sonnet, and Claude 3.5 Haiku. Available online: https://www.anthropic.com/news/3-5-models-and-computer-use.
25. Google DeepMind (2024, May 20). Gemini 2.5 Flash. Available online: https://deepmind.google/models/gemini/flash/.
26. Sahoo, P., Singh, A.K., Saha, S., Jain, V., Mondal, S., and Chadha, A. (2024). A systematic survey of prompt engineering in large language models: Techniques and applications. arXiv.10.1007/979-8-8688-0569-1_4 ↗
27. Language models are unsupervised multitask learners
28. Language models are few-shot learners
29. Chain-of-thought prompting elicits reasoning in large language models
30. Chia, Y.K., Chen, G., Tuan, L.A., Poria, S., and Bing, L. (2023). Contrastive chain-of-thought prompting. arXiv.
31. Li, B., Lee-Urban, S., Johnston, G., and Riedl, M. (2013, January 14–18). Story generation with crowdsourced plot graphs. Proceedings of the AAAI Conference on Artificial Intelligence, Bellevue, WA, USA.10.1609/aaai.v27i1.8649 ↗
32. Tang, B.J., Boggust, A., and Satyanarayan, A. (2023, January 9–14). Vistext: A benchmark for semantically rich chart captioning. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics, Toronto, ON, Canada.10.18653/v1/2023.acl-long.401 ↗
33. Kim, J., Lee, J., Choi, H.J., Hsu, T.Y., Huang, C.Y., Kim, S., Rossi, R., Yu, T., Giles, C.L., and Huang, T.H. (2025). Multi-LLM Collaborative Caption Generation in Scientific Documents. arXiv.10.1007/978-981-96-8912-5_6 ↗
34. Guo, S., Deng, C., Wen, Y., Chen, H., Chang, Y., and Wang, J. (2024, January 21–27). DS-agent: Automated data science by empowering large language models with case-based reasoning. Proceedings of the International Conference on Machine Learning, Vienna, Austria.
35. Michalowski, M., Wilk, S., Bauer, J.M., Carrier, M., Delluc, A., Le Gal, G., Wang, T.F., Siegal, D., and Michalowski, W. (2024, January 9–12). Manually-Curated Versus LLM-Generated Explanations for Complex Patient Cases: An Exploratory Study with Physicians. Proceedings of the International Conference on Artificial Intelligence in Medicine, Salt Lake City, UT, USA.10.1007/978-3-031-66535-6_33 ↗
36. Meta, A.I. (2025, April 05). The Llama 4 Herd: The Beginning of a New Era of Natively Multimodal AI Innovation. Available online: https://ai.meta.com/blog/llama-4-multimodal-intelligence/.
37. Understanding readability of large language models output: An empirical analysis10.1016/j.procs.2024.09.636 ↗
38. Smith, E.A., and Senter, R. (1967). Automated Readability Index, Aerospace Medical Research Laboratories, Aerospace Medical Division, Air Force Systems Command.
39. Chkirbene, Z., Hamila, R., Gouissem, A., and Devrim, U. (2024, January 3–5). Large language models (llm) in industry: A survey of applications, challenges, and trends. Proceedings of the 2024 IEEE 21st International Conference on Smart Communities: Improving Quality of Life using AI, Robotics and IoT (HONET), Doha, Qatar.10.1109/HONET63146.2024.10822885 ↗
40. Framework for LLM applications in manufacturing10.1016/j.mfglet.2024.09.030 ↗