XHABC: Enhanced Hybrid Artificial Bee Colony Algorithm for Recommending the Personalized Multi-Day Mixed Destination Tour

Developing a recommender system for multi-day tourist route planning that aligns with user preferences presents challenges in optimization. Many previous studies within the scope of Tourist Trip Design Problem (TTDP) have developed systems that recommend multi-day routes by utilizing the analogy of solving the Team Orienteering Problem (TOP). In addition, prior studies have not accommodated the user’s desire to visit a combination of several tourist categories, such as nature, shopping, and culinary, during a multi-day tour. We refer to this problem as a multi-day mixed destination tour. Therefore, in this research, we formulate the multi-day mixed destination tour problem as a Capacitated Team Orienteering Problem with Time Windows (CTOPTW) and introduce a new algorithm, XHABC, a hybridized version of Artificial Bee Colony (ABC) and Harris Hawk Optimization (HHO), to solve the problem. This study specifically focuses on combining two categories of Points of Interest (POIs), i.e., culinary and non-culinary POIs. To ensure the optimality of per-day routes, we develop an algorithm called DaySplitter based on the greedy strategy. In addition, we consider multi-attribute user preferences, which are integrated using the utility function in Multi-Attribute Utility Theory (MAUT). Based on the experimental results, XHABC consistently outperformed the comparison algorithms in both primary and secondary metrics, while additional evaluations further demonstrated its effectiveness, stability, and competitiveness in solving CTOPTW.