Dynamic bandwidth allocation in communication systems: an optimal control approach

The authors present the formulation and solution of an optimal bandwidth allocation strategy for virtual paths in the broadband integrated services digital network (ISDN) under nonstationary conditions. A state variable modeling approach is adopted to describe the dynamic behavior of the virtual path for different traffic classes. The problem is treated by Hamilton-Jacobi arguments, to formulate an optimal capacity allocation strategy for the equilibrium costate case. This is compared, using simulation, with the general solution of the problem. The present approach to the optimal capacity allocation problem differs from the majority of approaches in the literature on the optimal control of queueing systems which are based on the assumption of steady-state queueing models. The authors deal with the more pragmatic case of nonstationary network behaviour. i.e., they seek to optimize time-varying averages rather than steady-state averages