Multi-failure network restorability design in survivable transport networks

Abstract

The Dual Failure Restorability (DFR) problems involve the design of network topology to be restorable in the event of single and dual failures scenarios. We developed new integer linear programming (ILP) models to optimally design mesh topology networks with various survivability schemes; span restoration, p-cycle, DSP and path restoration to achieve any specified level of dual failure restorability in the networks. The first variation of the ILP models applies specified dual failure restorability limit to each pair of spans in the network, and the second applies the limit to average dual failure restorability in the entire network. We used 137 test-case networks, consisting of four network families; 10-node, 12-node, 15-node, and 18-node network families. The results show that the capacity cost increases as the specified levels of dual failure restorability increases, and the relative increase in capacity cost in sparsely connected networks is much higher compare to densely connected networks.