Personalized Travel Route Recommendation Using a Hybrid PSO–ACO Algorithm

With the rapid growth of smart tourism, personalized travel route recommendation has emerged as a critical challenge to enhance user experience and satisfaction. Traditional recommendation systems often fail to adequately capture users' dynamic preferences, sentimental inclinations, and the multifaceted attributes of Points of Interest (POIs). To address these limitations, we propose a hybrid PSO-ACO algorithm that leverages the global search capability of PSO and the local exploitation strength of ACO, while integrating user sentiment and POI similarity into the heuristic function.. The model integrates user sentiment derived from review texts and POI similarity based on rating matrices into an enhanced ACO heuristic function. Furthermore, a PSO module is employed to optimize the initial pheromone distribution of the ACO, thereby accelerating convergence and mitigating the risk of local optima. We validated our approach on a dataset comprising reviews and ratings from 593 tourists for 37 popular scenic spots within Beijing's Forbidden City. Experimental results demonstrate that the proposed PSO-ACO algorithm significantly outperforms benchmark methods, including the elite ant system and improved genetic algorithms, across key metrics of precision, recall, and F1-score. This work provides a robust and personalized framework for intelligent travel itinerary planning.