RUDN Mathematicians Found a Way to Accelerate the Modeling of Wireless Communication Systems

RUDN mathematicians studied a queueing system and developed a new algorithm for calculating the most efficient regime of server operation. The development may be used for modeling and further optimization of data transmission in wireless communication networks. The new algorithm will process calculations hundreds of times faster than its existing analogs. The results of the work were published in Applied Mathematical Modelling.

The theory of queues is used to optimize and model the work of modern telecom systems and communication networks. However, the majority of existing studies in this area are based on the assumption that queueing devices (e.g. CPUs) are 100% reliable. In reality this is often not true, as devices tend to malfunction. RUDN mathematicians considered a multiserver system in which a user can do a callback in case of a failure.

"We've analyzed a queueing system with callback and a quite complex process of operation. This system can be used to model actual ones that are not 100% reliable, i.e. may have failures in the course of their work. An example of such a process may be wireless transmission; if it malfunctions, a part of the data may be lost," said Sergey Dudin, a co-author of the work, and a junior research associate at the Institute of Applied Mathematics and Telecommunications, RUDN.

The theory of queues is based on the probability theory and mathematical statistics. It helps one select the most efficient queueing system and process structure taking into account queueing requirements that come in and out of the system, waiting time, and the length of queues. Several developments also account for possible device malfunctions but consider only complete breakdowns without paying attention to minor failures. In case of a glitch a current request is may be lost, or operations may start from the phase in which they stopped or be repeated from the very beginning. Although such systems are widely spread, they haven't been studied yet. RUDN mathematicians were the ones to lay down a theoretical basis for such studies.

Current optimization algorithms for such systems are quite slow. A team of researchers from RUDN developed an alternative algorithm that may be used to solve a wide range of problems and works several times faster than its analogs. In a one-processor system the new algorithm increases the speed of calculations 5 times compared to standard methodologies. An if there are 15 CPUs in the system, calculations are accelerated 170 times.

"The results of our work may be used to design new wireless communication networks and optimize existing ones. This may help reduce time and resources spent on design processes and improve service quality. For example, by selecting optimal system parameters we can reduce the chances of channel occupancy and increase the speed of mobile Internet connection," added Sergey Dudin.

More information:
Sergei Dudin et al. Retrial multi-server queuing system with PHF service time distribution as a model of a channel with unreliable transmission of information, Applied Mathematical Modelling (2018). DOI: 10.1016/j.apm.2018.09.005

Provided by RUDN University