A Parallel Evolutionary Algorithm for Multilayered Robust Network Design

Over the last decades multilayered networks have been a key design component for telecommunications infrastructure. Although some works have addressed the optimization of multi-overlay networks, just a few of them are contextualized for the current technological environment. This article focuses on solving the problem of designing a minimum-cost fault-tolerant multi-overlay network. the problem is NP-hard, and exact techniques are not appropriate to compute accurate solutions efficiently, so the application of sequential and parallel genetic algorithm to solve the problem is explored. the experimental analysis is performed on real-world scenarios, built from an IP/MPLS data network mounted over a DWDM optical network. the analysis shows that the studied genetic algorithms are able to obtain promising results, while the parallel model significantly speeds up the problem resolution.