ANFIS for Markovian Unreliable Retrial Queue with Differentiated Vacation

The performance analysis of an unreliable server Markovian retrial queueing system with balking and reneging of the customers has been proposed. The differentiated vacation policy has been considered which is a combination of \(k\) working vacation and complete vacation. To analyze the queueing system, Chapman-Kolmogorov equations are constructed and solved by implementing the matrix geometric method. The system metrics such as state probabilities, mean queue length, mean system length, mean waiting time in the queue, and total cost of the system are formulated. The sensitivity analysis of the system performance metrics has been done by taking illustrations. The hybrid soft computing approach Adaptive Neuro-Fuzzy Inference System (ANFIS) has also been implemented to explore the design of AI-based controller in order to control the traffic.