Research On Multi-Constraint QoS Routing Algorithms

Currently, lots of works have shown that the routing algorithms play an important role in achieving the network quality of service (QoS). QoS routing algorithm is one of the key issues in network researches. As the multi-constraint based network model can accurately reflect the actual QoS routing problems, the research on routing algorithms is of a great significance. In all the issues of multi-constraint routing algorithms, how to solve the DCLC (Delay-Constrained Least-Cost) problem is an important direction of recent network researches. All the existing routing algorithms have their own advantages on solving DCLC problem, however, how to solve the scalability of routing algorithms is a well-known problem.The QoS routing algorithms for solving the DCLC problem in hierarchical networks is mainly discussed in this thesis. Topology aggregation (TA) technique is often used to compress the vast amounts of network state information. It is an important way to solve the scalability problem of QoS routing algorithms with the application of the summarized information to compute the routing paths. Therefore, the accuracy of the TA algorithms to abstract the information plays an important role to solve the accuracy and feasibility problems of QoS routing algorithms. A new TA algorithm is presented in this thesis, and it has been verified to be feasible and effective by lots of simulation experiments. On TA basis, through extensive simulation experiments, the two time-scale Markov decision process based QoS routing algorithm has been proved to be effective in solving the DCLC problem in hierarchical networks. Furthermore, this new QoS routing algorithm is also an outperforming way in solving the DCLC problem compared to other existing QoS routing algorithms.