Development of a convolutional algorithm for calculating the stationary characteristics of resource loss systems with requirements-depending serving time

Resource loss systems (ReLS) are often used to analyze and calculate the performance metrics of wireless communication systems. The authors consider a multiserver ReLS with discrete requirements and service times that depend on the volume required resources, which allows one to analyze the efficiency of elastic traffic transmission in wireless networks. The evaluation of probabilistic characteristics for the model is complicated by the need to perform multiple convolutions of resource requirement distributions. To reduce the complexity of the calculations, the paper proposes a convolutional algorithm that allows one to obtain probabilistic measures, such as the loss probability and the average volume of occupied resources, during the calculation of normalization constant. The numerical analysis of the loss probability and the average volume of occupied resources for discrete distributions of resource requirements is carried out.