Design and Improvement of a Large Language Model Travel Planning System

With the rapid advancement and widespread adoption of large language models (LLMs), an increasing number of individuals have incorporated these modelsinto their daily routines. To address persistent challenges in travel planning, this paper proposes an LLM-powered specialized travel planning system. By processing concise yet essential user inputs, the system can generate comprehensive travel plans tailored to individual requirements. This study designs user input templates and evaluates the response quality of the large language models Deepseek and Kimi. The responses are assessed and compared using predefined metrics to establish a preliminary system framework and identify limitations, such as insufficient information input, the computational complexity of evaluation metrics, and inefficient and deficient information output. This paper provides potential solutions that include API (application programming interface) integration, enhanced user-system interaction, optimized prompt engineering, and implementation of advanced algorithms. This paper systematically reviews the system design, identifies prevailing challenges, and outlines future development trajectories. It validates the feasibility and potential of the Large Language Model Travel Planning System while establishing a comprehensive framework for LLM-based planning research, serving as a valuable reference for researchers in related fields.