A survey of AI-driven personalization for high performance group travel recommender systems

Abstract

Travel Recommender Systems (TRS) have become an essential part of modern digital tourism. TRS is designed to assist travelers in planning trips by filtering vast amounts of travel data to provide personalized suggestions. Advances in artificial intelligence, particularly in machine learning, optimization, neural embeddings, and personality-based modeling, have significantly improved the ability of these systems to adapt to user preferences, contextual factors, environmental conditions, and behavioral patterns. While individual-based recommender systems are relatively mature, supporting collaborative travel planning remains a complex challenge. This survey explores the state of the art in Group Travel Recommender Systems (GTRS). It analyzes preference aggregation strategies, algorithmic approaches, and existing methods that address cold-start issues. Furthermore, it reviews evaluation models used for assessing GTRS performance. The findings reveal challenges related to fairness, negotiation support, scalability, real-time adaptability, transparency, and system integration. The study also identifies key research gaps and emphasizes the growing need for more intelligent, context-aware, and socially adaptive GTRS.