GenAI job scheduling system for solving a flexible job shop scheduling problem

Abstract Generative artificial intelligence (GenAI) applications in job scheduling are expected to help schedulers embed their requirements into scheduling models in a more user-friendly way to generate customized scheduling results. However, there are still very few such applications, while using existing general-purpose GenAI services is inconvenient and prone to data leakage risks. To solve these problems, this study established a GenAI job scheduling system. By hosting the GenAI job scheduling system locally, schedulers can avoid the leakage of order- or recipe-related information that may occur when uploading to the cloud-based GenAI service. In the GenAI job scheduling system, a user interface is designed for users to enter queries in natural language. The user’s query is then analyzed to extract his/her requirements related to the scheduling task, thereby building an extended three-field notation (ETFN) of the scheduling problem. A customized genetic algorithm (GA) is generated to help solve the mathematical programming (MP) model corresponding to the ETFN, thereby updating invalid code or adding new code to the basic GA application. The effectiveness of the GenAI job scheduling system has been tested in a flexible job shop case.