Dynamic Communication Networks with Capacity Constraints

Communication networks with link transit times are modeled by linear graphs with branch time delays and finite branch capacities. Memoryless linear routing strategies as well as linear routing strategies with finite memory are defined. The state reachability problem in both cases is considered, and the sets of reachable demand vectors are exhibited. The problem of finding optimal routings which minimize network losses subject to demand constraints is formulated as a linear program, and extensions to infinite memory and time-varying systems are given.