This research will examine heterogeneous telecommunication networks, and develop novel solutions techniques for designing these networks to survive under link or node failure. The interconnections of Local Area Networks (LANs), the connection of LANs to Wide Area Networks (WANs), the interconnection of WANs, are all examples of heterogeneous data networks. Each of these networking problems have unique characteristics based on the communications protocols and the routing behavior of the interconnect devices. The research will consider the design problem for heterogeneous networks in the framework of formulating a cost minimizing mathematical programming (or optimization) approach. Survivable data networks are a special class of minimum-weight k-connected capacitated spanning networks. Designing minimum-weight k-connected spanning networks is NP-hard. The problems that this research addresses involve the consideration of additional constraints, and are also NP-hard. The solution techniques are based on a hybrid of simulated annealing and graph theory. In addition to successfully implementing these methods for some heterogeneous network design problems, the PIs propose to work on the design of international heterogeneous telecommunications networks.
