Enhancing Tourism Recommender Systems for Sustainable City Trips Using Retrieval-Augmented Generation

TL;DR

The paper addresses the challenge of making Tourism Recommender Systems sustainable by integrating a sustainability metric into a Retrieval-Augmented Generation pipeline. It constructs a Wikivoyage-based knowledge base for 160 European cities and defines the S-Fairness indicator as $\psi(c_i^j) = 0.334 \cdot \rho(c_i) + 0.385 \cdot \sigma(c_i^j)$ to balance city popularity and seasonality during prompt augmentation. The SAR approach is evaluated against a Baseline across two open-source LLMs (Llama-3.1-Instruct-8B and Mistral-Instruct-7B), using synthetic test cases to show SAR generally matches or improves performance while reducing hallucinations. The work demonstrates that explicit sustainability information can guide multi-stakeholder recommendations toward less crowded, more responsible travel options, with practical implications for TRS design and deployment; future work should expand data sources and metrics and explore few-shot and conversational capabilities.

Abstract

Tourism Recommender Systems (TRS) have traditionally focused on providing personalized travel suggestions, often prioritizing user preferences without considering broader sustainability goals. Integrating sustainability into TRS has become essential with the increasing need to balance environmental impact, local community interests, and visitor satisfaction. This paper proposes a novel approach to enhancing TRS for sustainable city trips using Large Language Models (LLMs) and a modified Retrieval-Augmented Generation (RAG) pipeline. We enhance the traditional RAG system by incorporating a sustainability metric based on a city's popularity and seasonal demand during the prompt augmentation phase. This modification, called Sustainability Augmented Reranking (SAR), ensures the system's recommendations align with sustainability goals. Evaluations using popular open-source LLMs, such as Llama-3.1-Instruct-8B and Mistral-Instruct-7B, demonstrate that the SAR-enhanced approach consistently matches or outperforms the baseline (without SAR) across most metrics, highlighting the benefits of incorporating sustainability into TRS.

Figures (2)

• Figure 1: Pipeline of our proposed modified RAG system. The green dotted box shows the modified augmentation phase with the addition of the sustainability metric.
• Figure 2: Exploratory Data Analysis for Wikivoyage Articles and Listings