AGENT-BASED SIMULATION ENHANCED BY MULTILAYER PERCEPTRON FOR MANAGEMENT INFORMATION SYSTEMS: AN APPLICATION ON OPEN EDUCATION SYSTEM

This study develops and applies a modeling approach that combines a multilayer perceptron (MLP) and agent-based simulation (ABS) to an open and distance education system, aiming to decrease dropout rates and enhance system sustainability. The approach incorporates a management information system that tests various scenarios and decisions. Designed for similar social systems, this research provides unique methodologies and data that could serve as a model for related studies. The MLP model used in the study, accurately predicts students' grade point averages, enabling to integration into ABS models for more precise simulations. Administrative strategies successfully increase active students and graduates in the short term, despite an initial reduction and subsequent fluctuation in dropout rates. External economic factors negatively affect student engagement and success, though the system shows resilience, adapting over time. The study also simulates scenarios, showcasing the management system's capabilities emphasizing the system's broad applicability and effectiveness.