Resource Distribution and Balancing Flows in a Multiuser Network

Balancing strategies for control flows and resources are studied in computational experiments on a mathematical model of a multiuser network. An algorithmic scheme for sequentially solving a chain of lexicographically ordered problems of finding nondiscriminating the maximin distributions of flows is proposed. A procedure for step-by-step equalizing the splitting of internodal flows along all the existing shortest paths is developed. At each iteration, a part of the available resource is distributed equally among all pairs of correspondent nodes for which there is a possibility of flow transmission. The results obtained in the course of the experiments make it possible to trace the dynamics of changes in the capacity of the edges, up to reaching the maximum network load. Bottlenecks in various networks are analyzed and compared with a monotonic increase in the flows. Special diagrams are given.