Response Time Prediction of M/M/1SRPT Queuing System Using Simulation Modeling and Artificial Intelligence

In queueing systems, users need to know the expected response time of their jobs for decision-making and proofing system reliability. Because of this reason, there is a need to predict the response time of given jobs when a specific discipline is implemented in the queuing system. In this paper, we proposed a novel method combining simulation modeling and artificial intelligence methods to predict job response time on the M/M/1/SRPT queue. Simulation modeling is used for generating data, which is then used by the artificial methods to do the response time prediction. In our proposed approach, three attributes are used to predict the response time: job processing time, total processing time in the system, and total processing time of the preceding jobs in the queue. These attributes are used in both artificial intelligence methods: linear and support vector regression (SVR). Based on the case study testing result, our proposed method resulted in an average variance score of 94.5% using linear regression, 99.7% using SVR polynomial, and 99.8% using SVR RBF, which proves the prediction accuracy.