Abstract
The tourism industry is a significant contributor to the global economy, responsible for generating nearly 10% of the world’s GDP and employing around 9% of the global workforce. A crucial aspect of this industry is personalised itinerary recommendation, where visitors’ preferences and constraints are taken into account to create customised travel plans. This task involves selecting the best points of interests (POIs) for visitors in various cities and then schedule these POIs as an itinerary considering numerous constraints. However, due to the varied ways in which researchers have defined the itinerary recommendations, it can be challenging for new researchers to locate up-to-date literature on the topic. As a result, this paper aims to review existing research in this area and provide a taxonomy of the works based on problem formulations, proposed techniques, constraints, and features used. We divide the study into two directions: user satisfaction and provider satisfaction, where user satisfaction is derived non-personalised and personalised POI/ Itinerary recommendations. We also discuss the data sources, techniques ranging from optimisation approaches to deep learning and evaluation methodologies commonly used in this field. Finally, we highlight the importance of personalised itinerary recommendation and identify areas for future research to address the current challenges.
•We discuss deep learning techniques based POI/itinerary recommendations.•It focuses two research directions: user satisfaction and provider satisfaction.•It explores user satisfaction into non-personalised and personalised sub-categories.•We utilise provider satisfaction based on the utilisation of their resources.•We emphasis the sub-areas of operations research, recommendations and fairness.•We also describe some significant research directions for the future researchers.