Optimizing reliable network topology design using dynamic programming

Abstract

This thesis addresses four reliable network topology design problems that consider reliability, cost, and/or bandwidth performance metrics. The optimization problems include one or two objectives and one constraint, each of which is known NP-hard. All problems consider all-terminal and/or two-terminal reliability measures. Four heuristic dynamic programming approaches are proposed, and 10 order criteria are used to improve their optimality and time efficiency. Simulations on hundreds of networks show the merits of the proposed methods.