Labour Optimization and Staff Scheduling Using LSTM and XGBoost with Real-Time Alert System

Abstract

Workforce scheduling is a complex task in industries with fluctuating staffing demands, such as healthcare, hospitality, and logistics. Traditional approaches rely on manual planning or static rules, which often lead to inefficiencies and employee dissatisfaction. In this work, we propose an AI-driven labour optimization and staff scheduling system that integrates a hybrid LSTM–XGBoost forecasting model with an optimization engine for fair and compliant shift generation. To enhance usability, the system includes a chatbot for natural language interaction and an alert mechanism for real-time schedule notifications. Experimental results show that the hybrid model reduces prediction error by over 20%, schedules are generated in under five seconds for more than fifty employees, and alert systems achieved 92% query accuracy and 100% delivery reliability, respectively. These results demonstrate the system’s potential for practical deployment in modern workforce management.

Keywords