Key points are not available for this paper at this time.
ABSTRACT Water resources management is inherently complex and dynamic, characterized by strong time-sensitivity and uncertainty. Traditional water resources information systems rely on fixed-process decision-making, which often lack flexibility to accommodate emerging needs and are insufficiently responsive to sudden events or environmental changes. To overcome these limitations, this paper proposes an Intelligent Water Resources Management System (IWMS-LLM) based on Large Language Models (LLMs), and introduces a visual workflow orchestration approach tailored to water resources management tasks. The system leverages LLMs as its core engine to support workflow orchestration and drive task execution, thereby establishing a problem-driven service model. IWMS-LLM comprises four key modules: (1) ingestion, preprocessing, and management of multi-source data; (2) encapsulation of domain models and operational process components; (3) visual workflow orchestration for task configuration; and (4) intelligent task execution via natural language processing. IWMS-LLM performs well in terms of adaptability to diverse scenarios and system scalability. It effectively meets diverse needs in water resources management, lowers the barriers to development and utilization, and provides an innovative technological pathway for the intelligent advancement of water resources management systems.
Guo et al. (Sat,) studied this question.