ANALYSIS OF APPROACHES TO MODELING QUEUING SYSTEMS

The queuing theory has a wide practical application, therefore, the optimization of the queuing systems (QS) has a signifi-cant role in the modern world. To simulate and analyze the characteristics of the queue on the GPSS simulation environment can be used. Its shortcomings, such as closed source code, requirements for system compatibility with the environment, lack of API and others, lead to problems of its application when developing your own programs using QS. Three solutions have been proposed that make it possible to simulate a single-channel QS; they are devoid of the listed disadvantages, but they require more time for modeling. Two solutions (implementations in GNU Octave and in Python), based on the charts given in the work, are convenient for modeling and analyzing QS, since their operation time does not depend on the mean value of the distribution, or on the number of applications coming into operation. The third solution, for which two Python implementations are described, in both cases shows a linear dependence of the execution time due to the processing of requests in real time. This approach allows us to obtain a practical time for processing requests without taking into account the costs of creating and deleting them, which improves the quality of comparison of the characteristics of the queue with the theoretical ones. At the same time, such a solution can be implemented into existing real-time QS in order to obtain the characteristics of the queue. It is also shown that the practical values obtained as a result of the work of the programs converge to the theoretical ones.

Authors: M. O. Dobrokhvalov, P. V. Korytov, S. I. Stepanova, A. A. Tarasova, Ar. Yu. Filatov

Direction: Informatics, Computer Technologies And Control

Keywords: QS, queuing system, GPSS, queue, simulation

---

View full article