Qpgs: An Intelligent, Offline-capable Web-based Question Paper Generation System For Outcome-based Education In Higher Educational Institutions

Abstract

The Manual Preparation Of Examination Question Papers In Higher Educational Institutions Is A Labor-intensive, Error-prone, And Time-consuming Activity. Faculty Members Often Spend Between Two And Four Hours Assembling A Single Question Paper, Ensuring Appropriate Difficulty Distribution, Compliance With Examination Templates, And Alignment With Course Outcomes Defined Under Outcome-Based Education (OBE) Frameworks. This Paper Presents QPGS—a Question Paper Generation System—a Cloud-native, Offline-capable, And Role-based Web Application Designed To Automate And Streamline Examination Paper Creation For Institutions Affiliated With The Savitribai Phule Pune University (SPPU) Pattern. The System Employs A React 19-based Single-page Application (SPA) With Firebase Firestore As A NoSQL Backend, Firebase Authentication For Secure Role-based Access, And An IndexedDB-based Offline Queue For Uninterrupted Operation In Low-connectivity Environments. At The Core Of Paper Generation Lies A Configurable OR-group Template Engine Combined With The Fisher-Yates Shuffle Algorithm, Ensuring Randomized Yet Difficulty-balanced Question Selection. Additionally, QPGS Integrates An AI-powered Question Generation Pipeline Capable Of Extracting Syllabus Content From Uploaded PDF And DOCX Files And Generating Pedagogically Appropriate Questions. Evaluation Of The System In A Pilot Institutional Setting Demonstrated An 80% Reduction In Paper Preparation Time, With Successful Generation Of Well-balanced Papers In Under Two Minutes. The System's Approval Workflow, LaTeX Rendering Support, And Analytics Dashboard Further Contribute To Its Academic Value. A System Usability Scale Score Of 82.4 Confirms Excellent Usability Across Non-technical Faculty Users

Keywords