| In this Thesis, we document the results of our investigation of four important problems in the domain of telecommunication network routing and traffic engineering. The techniques we use involve stochastic learning and Learning Automata (LA).; Our first contribution consists of two efficient solutions for maintaining single-source shortest path routing trees in networks, where the weights of the links connecting the nodes of the network change continuously in a random manner following an unknown stochastic distribution.; In the second problem, we were interested in maintaining shortest paths between all pairs of nodes in a dynamically changing network. Again, for this problem, we have proposed two LA-based efficient solutions.; Our third contribution was in the area of QoS routing and traffic engineering in networks. More specifically, we have proposed an adaptive online routing algorithm that computes bandwidth-guaranteed paths in MPLS-based networks, using a Random-Race-based learning scheme that computes an optimal ordering of routes.; The last problem that we studied was that of designing routing schemes that would successfully operate in the presence of adversarial environments in Mobile Ad Hoc Networks (MANETs). The need for fault tolerant routing protocols for MANETS was identified recently by Xue and Nahrstedt, and in our research, we have proposed a new fault-tolerant routing scheme that uses a stochastic learning-based weak estimation procedure.; All our proposed solutions have been demonstrated to be superior to the state-of-the-art. |
