Strategic and tactical models for multi-destination traffic routing in telecommunication networks

This thesis is dedicated to the development of efficient traffic routing strategies and related performance evaluation models for multi-destination traffic in telecommunication networks. Multi-destination traffic refers to a traffic type that can be routed to any one of a multiple number of destinations (e.g., an 800 number with multiple answering points). While multi-destination applications are becoming more pervasive, the literature is devoid of papers dealing with multi-destination traffic.;In addition to dealing with the topic of multi-destination traffic, the thesis first presents several strategic optimization models which can be used to optimally allocate multi-destination traffic load to destination links under a proportional routing scheme from a steady-state analysis standpoint. Here, we provide models in which optimal decision parameters are determined from three different perspectives: (i) balanced link utilization, (ii) cost optimization and (iii) both combination of the foregoing methods. Indeed, a balanced link utilization approach has been implemented in a real network application and a very noticeable improvement has been observed.;Additionally, several tactical models are presented. These models, which are based on the Erlang fixed-point approach, are formulated to evaluate the network blocking probability of a number of multi-destination traffic routing strategies. The presented models have been shown to be very accurate and to converge for all the empirical analyses made to date. The models capture the state-dependent dynamics for both single and multi-destination traffic in an end-to-end network setting. State-dependent dynamic routing strategies are shown to be superior to the static routing strategy for an end-to-end network involving multi-destination traffic. Further, it is shown that a better performance is achieved with a routing strategy that employs more real-time network information in an end-to-end network setting, even though it is more complex to model and implement such a routing strategy.;In order to assess the efficiency of the proposed tactical routing strategies, we developed a max-flow bound. In this manner, it is shown that our dynamic strategies are very efficient in successfully routing multi-destination traffic in an end-to-end network.;Finally, several empirical analyses and a real-world application are presented. It is clear from the presented results that the strategic and tactical models developed herein do indeed yield efficient routing strategies for multi-destination traffic.