AI-Powered Dynamic Timetable Scheduler Architecture

Abstract

Automated timetable scheduling in educational institutions remains a complex combinatorial optimization problem. Traditional manual scheduling approaches are time-consuming, error-prone, and often fail to accommodate institutional constraints and faculty preferences. This paper presents an AI-powered dynamic timetable scheduler that integrates a React and Tailwind CSS frontend with a FastAPI backend powered by Google OR-Tools' CP-SAT (Constraint Programming Satisfaction) solver. The proposed system enables administrators to dynamically configure institutional parameters—including departments, classes, subjects, rooms, faculty members, and their availability constraints— without hardcoded credentials or static data. The backend formulates inputs as a constraint satisfaction problem (CSP) and computes optimal multi-class timetables across all sections. Our architecture emphasizes scalability for large datasets, responsive performance, security through database-managed authentication, and a polished user interface with real-time constraint handling. The system successfully demonstrates automatic scheduling with integrated constraint handling, dynamic faculty availability management, and room allocation across multiple departments and class sections.