Waiting Time Prediction in Queue Management: Leveraging Machine Learning Approach

Ahstract- This research highlights a novel approach to queue management systems (QMS), offering a scalable, cost-effective, and intelligent solution to contemporary challenges. The pro-posed study evaluates regression algorithms along with Long Short Term Memory (LSTM) model to enable real-time tracking and estimation, considering the continuous nature of the pre-diction of the waiting time required by a client before taking a service. Among the models evaluated for waiting time prediction, the Voting Regressor(VR) demonstrates superior performance across simulation scenarios, achieving the highest R2 score of 0.944 and the lowest mean absolute error (MAE) of 1.828 minutes. The QMS dynamically allocates customer requests to available service counters, optimizing resource utilization and improving operational efficiency. This work addresses the limitations of traditional physical systems, such as the need for physical presence, inefficient space and time utilization, and limited scalability.