Large Language Models for Decision-Making Enhancement in Tourism Package Customization

Abstract

The tourism industry faces significant challenges in providing personalized travel recommendations that cater to diverse traveler preferences and needs. This paper presents an approach utilizing Large Language Models (LLMs), specifically the open-source LLaMA model, to enhance decision-making in tourism package recommendations. By integrating a comprehensive database of attractions, amenities, and activities, the proposed system generates tailored suggestions based on user profiles, including demographics, interests, and travel duration. The application of few-shot learning techniques in prompt engineering allows the model to effectively interpret user inputs and produce relevant recommendations with minimal examples. This research aims to demonstrate how LLMs can serve as powerful decision aids in logistics within the tourism sector, ultimately improving user satisfaction and streamlining the travel planning process.