Executive Summary : | The project focuses on profit and cost investigations of finite and infinite Markov and non-Markov queueing models with retrial orbit under admission control policies. These models play a significant role in communication systems, manufacturing processes, production, and machine repair problems. Customers join a virtual waiting place called retrial orbit when they find an unavailable or busy server, hoping that the server becomes available to facilitate them. Multiple orbits (double and triple orbit) queues are also incorporated, providing customers more options to select the retrial orbit based on their paying capacity. Admission control policies like F-policy, p-policy, and (F-p)-policy are crucial features of queueing models that control customer congestion. Implementing these policies in the proposed Markov and non-Markov retrial queueing models can enhance service quality, enhance profit, and reduce costs in real-world queueing scenarios. Admission control parameters help customers decide whether to queue up for desired services based on queue length or total waiting time. Optimal joining strategies of customers in unobservable and observable queues will be studied, providing valuable insights for system designers and decision-makers to reduce system costs and congestion problems. The cost-reward function will be framed and analyzed based on arriving customers' joining strategies. several mathematical techniques, such as recursive, probability generating, matrix analytic, and Laplace-stieltjes transform, will be used to solve non-Markov retrial models. soft computing-based algorithms, such as the harmony search algorithm, genetic algorithm, and sine-cosine method, will be implemented to minimize the system's total cost, helping system managers formulate business policies and strategies. |