A Data-Driven Framework for Group-Aware Itinerary Recommendation
MSc Thesis Defense by: Mobin Ali Momin
Date: Wednesday, July 22th, 2026
Time: 2:30pm to 3:30pm
Location: Online (MS Teams)
Abstract:
Travel itinerary planning can be seen as a constrained recommendation problem in which a sequence of points of interest must reflect user preferences while remaining feasible under practical travel conditions. The challenge becomes more complex for group travel because members may have different interests but must generally follow one shared itinerary. A successful group itinerary should therefore balance collective relevance, individual satisfaction, fairness, and route feasibility.
This thesis proposes the Group-Aware Intention-Based Itinerary Recommendation System (G-AIR), which uses a deep learning approach to generate feasible itineraries for groups of users. The model learns candidate POI relevance from historical travel behavior and combines it with group preference aggregation and constraint-aware itinerary construction. Travel time, queue time, visit duration, group support, and fairness are incorporated when selecting and ordering POIs. The model also supports a controlled optional activity choice that allows temporary subgroup flexibility while preserving the overall shared itinerary.
The proposed approach was evaluated on multiple tourism datasets and compared with established itinerary recommendation baselines. The results show that G-AIR provides competitive recommendation accuracy while producing feasible itineraries that effectively use the available travel time. The group-aware planning process also supports balanced satisfaction across members and adapts to different tourism environments. These findings demonstrate that combining learned POI relevance with group-aware and constraint-aware planning can provide a practical approach to group itinerary recommendation.
Keywords: Group-Aware Itinerary Recommendation, Intention-Aware Recommendation, Graph Convolutional Networks (GCN), Deep Learning
Thesis Committee:
Internal Reader: Dr. Muhammad Asaduzzaman
Internal Reader: Dr. Saeed Samet
Advisor: Dr. Pooya Moradian Zadeh
Chair: Dr. Ikjot Saini
