The rapid evolution of digital tourism platforms has demonstrated the pressing need for automated systems that can transform fragmented travel information into cohesive and highly personalized itineraries. While traditional systems offer isolated recommendations, users increasingly expect detailed day- wise plans that reflect real travel behavior, practicality, and comfort. The complexity of modern travel decision- making—shaped by cost fluctuations, diverse preferences, and varying destination characteristics— demands solutions capable of synthesizing multiple factors simultaneously. The system discussed in this research addresses these expectations by structuring user requirements into logical sequences that mirror the planning style of experienced human travel coordinators. This enhanced capability results in itineraries that feel both intuitive and realistic, bridging the gap between raw data and meaningful travel experiences.Unlike conventional recommendation platforms that provide isolated suggestions, the system emphasizes the creation of complete, day-wise itineraries that reflect realistic travel behavior and geographical flow. It aims to reduce planning fatigue and improve decision-making efficiency by ensuring that suggested activities align with human comfort levels, time availability, and practical feasibility. The methodology prioritizes organization, clarity, and sequencing, ensuring that travelers receive itineraries that not only highlight important attractions but also provide a balanced distribution of activities across the entire trip.
Keywords
Artificial Intelligence, Large Language Models, Travel Recommendation Systems, Personalization, FastAPI, Generative AI.
Conclusion
This research presented a novel approach to automated travel planning by integrating Large Language Models (LLMs) with a structured application layer. The developed system successfully addresses the limitations of traditional recommendation engines—which often provide fragmented lists—by synthesizing cohesive, day-wise itineraries that respect user constraints such as budget, interest, and pace. By leveraging the semantic reasoning capabilities of Google’s Gemini model and the speed of FastAPI, the system achieves a balance between creative personalization and logical feasibility.
The experimental results demonstrated a 92% Constraint Satisfaction Rate and a high degree of logistical accuracy, validating the efficacy of the prompt engineering strategies employed. The system effectively mitigates common LLM issues, such as hallucinations, through strict schema enforcement and context injection. It offers a scalable solution that reduces the cognitive load on travelers, transforming hours of manual research into a near- instantaneous, actionable plan.
However, the study also identified limitations regarding real- time data accuracy, specifically concerning dynamic pricing and variable opening hours. Future work will focus on bridging this gap by integrating real-time APIs (such as Google Maps Platform or Amadeus) to validate the LLM’s suggestions against live data. Additionally, we aim to expand the system to support multi-modal outputs, providing users with map visualizations and image galleries alongside their text-based itineraries, further enhancing the user experience in the smart tourism ecosystem.
References
[1]D. Gavalas, C. Konstantopoulos, K. Mastakas, and G. Pantziou, “Mobile recommender systems in tourism,” Journal of Network and Computer Applications, vol. 39, pp. 319–333, 2014.
[2]U. Gretzel, “Intelligent systems in tourism: A social science perspective,” Annals of Tourism Research, vol. 38, no. 3, pp. 757-779, 2011.
[3]S. Li, J. Cao, and V. C. S. Lee, “A Comprehensive Survey on Large Language Models for Recommendation,” arXiv preprint arXiv:2305.06999, 2023.
[4]H. Liu, et al., “LLM-Rec: Personalized Recommendation via Prompting Large Language Models,” arXiv preprint arXiv:2307.15780, 2023.
[5]
M. Nilashi, et al., “Travelers decision making using online review in social network sites: A fuzzy-logic approach,” Tourism Management, vol. 59, pp. 233-242, 2017.
[6]Google AI, “Gemini API Documentation,” [Online]. Available: https://ai.google.dev/docs. [Accessed: Nov. 20, 2025].
[7]F. Ricci, “Travel Recommender Systems,” in IEEE Intelligent Systems, vol. 17, no. 6, pp. 55-57, 2002.
[8]J. Borràs, A. Moreno, and A. Valls, “Intelligent tourism recommender systems: A survey,” Expert Systems with Applications, vol. 41, no. 16,
pp. 7370-7389, 2014.
[9]A. Vaswani et al., “Attention is all you need,” in Advances in Neural Information Processing Systems, vol. 30, 2017.
[10]D. Buhalis and R. Law, “Progress in information technology and tourism management: 20 years on and 10 years after the Internet—The state of eTourism research,” Tourism Management, vol. 29, no. 4, pp. 609-623, 2008.
[11]G. Adomavicius and A. Tuzhilin, “Toward the next generation of recommender systems: A survey of the state-of-the-art and possible extensions,” IEEE Transactions on Knowledge and Data Engineering, vol. 17, no. 6, pp. 734-749, 2005.
[12]Z. Xiang and D. R. Fesenmaier, “Big data analytics, tourism design and smart tourism,” Tourism Geographies, vol. 19, no. 1, pp. 107-124, 2017.
[13]H. Werthner and F. Ricci, “E-commerce and tourism,”
Communications of the ACM, vol. 47, no. 12, pp. 101-105, 2004.
[14]K. Kabassi, “Personalization systems for tourism: a comparative study,” User Modeling and User-Adapted Interaction, vol. 20, no. 1,
pp. 83-118, 2010.
[15]T. B. Brown et al., “Language models are few-shot learners,” arXiv preprint arXiv:2005.14165, 2020.
[16]B. Pan, T. MacLaurin, and J. C. Crotts, “Travel blogs and the implications for destination marketing,” Journal of Travel Research, vol. 46, no. 1, pp. 35-45, 2007.
[17]S. Zhang, L. Yao, A. Sun, and Y. Tay, “Deep learning based recommender system: A survey and new perspectives,” ACM Computing Surveys (CSUR), vol. 52, no. 1, pp. 1-38, 2019.
[18]D. Wang, S. Park, and D. R. Fesenmaier, “The role of smartphones in mediating the touristic experience,” Journal of Travel Research, vol. 51, no. 4, pp. 371-387, 2012.
[19]OpenAI, “GPT-4 Technical Report,” arXiv preprint arXiv:2303.08774, 2023.
H. Lin, Y. Fan, and P. J. Zhang, “A Deep Learning Approach for Next Point-of-Interest Prediction in Tourism,” IEEE Access, vol. 8, pp. 18933-18943, 2020.
How to Cite This Paper
Himani Rajput, Prabhat Yadav, Vishal Tiwari, Prachi Verma (2025). Personalized Travel Itinerary Generation Using Large Language Models and Generative AI. International Journal of Computer Techniques, 12(6). ISSN: 2394-2231.
