A Hybrid Approach Integrating User Preferences and POI Popularity for One-Day Travel Itinerary

The Tourist Travel Design Problem (TTDP) poses a significant challenge in crafting itineraries for tourist’s keen on exploring multiple Points of Interest (POIs) within a city while adhering to individual travel constraints. This study introduces a novel hybrid evolutionary approach that combines genetic algorithms and simulated annealing to automate the travel planning process. Empirical findings on real-world datasets illustrate the efficacy of this approach in addressing diverse optimization objectives and generating a varied array of travel itineraries for tourists, thereby confirming its scalability across different problem sizes and optimization goals. Furthermore, the research underscores the importance of personalized visit durations achieved through tailored travel recommendation algorithms, with Memtic-0 algorithms proving particularly effective in prioritizing trip popularity and user interest to enhance overall trip quality and satisfaction. Leveraging simulated annealing, the hybrid algorithm demonstrates superior performance compared to alternative methods such as 2-Opt and 3-Opt, thanks to its stochastic nature facilitating escape from local optima. Optimization through parameterization based on city constraints further enhances convergence speed and effectiveness, emphasizing the critical role of personalized trip duration and efficient convergence strategies in developing robust travel recommendation systems that benefit both users and tourism stakeholders.