A hybridized ANFIS-ANN based queueing model with orbital search, service interruptions during working vacation

Queueing models are widely used to analyze and optimize service systems in which customer arrivals, service times, and system dynamics are stochastic. This paper presents an unreliable Markovian single-server retrial queueing model with impatient customers and breakdowns, under a working-vacation policy. The proposed model accounts for various server states, including busy, breakdown, orbital search, and working vacation. Customers arrive according to a Poisson process and may either enter service, join an orbit, or abandon the system due to impatience. The server operates under an adaptive vacation policy, offering reduced service during vacations, and is subject to breakdowns that require repair. To improve predictive accuracy and system performance evaluation, we employ adaptive neuro-fuzzy inference system and artificial neural network models. These intelligent computational techniques provide robust approximations for system parameters, enhancing decision-making in queue management. The effectiveness of the proposed model is demonstrated through numerical simulations, offering valuable insights into the impact of system parameters on overall performance.