Models for parameter optimization of Timed Token Rotation Protocols

In this paper, models are presented with whose help it is possible to undertake parameter optimization of Timed Token Rotation Protocol-systems. The parameters to be optimized are the waiting time of one (or more) station(s) and the throughput. The basis for these models are, on the one hand, approximations of the waiting time achieved through analytic heuristic means, which were validated by simulation, and on the other hand, modern methods of nonliner optimization. The approximation formula for the waiting time is represented as the minimizing objective function, wherein the user requirements that are to be met in the search for an optimum are taken into consideration in the form of constraints (here: inequalities). Introduced in this paper are mono-criterium operation, where only the ‘waiting time’ criterium is considered, and bi-criteria operation, where the goals are both minimization of waiting time and maximization of throughput. Finally, the introduced optimization models are applied to examples.