GORA: an algorithm for designing optical cross-connect nodes with improved dependability

This paper considers the problem of enhancing the dependability (availability and its in¯uencing factors) of optical cross connect nodes. These nodes are expected to be a crucial part of future optical communication networks. The genetic optimum redundancy allocation (GORA) algorithm proposed here is based on a hybrid genetic algorithm and solves the dependability enhancement problem by optimizing the allocation of redundant modules in the nodes. Three redundancy schemes are considered: (1) global duplicate (2) local standby and (3) k-out-of-n:G redundancy. After applying the GORA algorithm to four example nodes, the optimized third redundancy scheme provided the overall best dependability performance. This suggests that redundancy is more effective when provided at the OXN subsystem level. Further capabilities of the GORA algorithm as a network planning tool also presented and some improvements are proposed. q 2002 Elsevier Science B.V. All rights reserved.