In this paper, we propose a novel integrated approach for designing survivable off-line Elastic Optical Network (EON) to minimize energy consumption and spectrum requirement. Single link failure recovery is ensured by applying multipath based routing and spectrum allocation (RSA). The RSA problem to minimize spectrum utilization is a NP-Hard one. Integer Linear Programming model is formulated for survivable multipath RSA (SM-ILP) to obtain the optimal solution for a small sized network. As the solution of SM-ILP gets intractable with the growth of the network size, two heuristics based on Greedy Algorithm (SM-GR) and Genetic Algorithm (SM-GA) are also developed to obtain near optimal solution for large-sized networks. Simulation studies have been carried out to evaluate the performance of the heuristics, and the results are compared with that obtained by solving the ILP formulation for small problem size. It is observed that for the considered simulation set up, proposed heuristics provide solution with variation of at most 6.77% for energy consumption and 3.71% for spectrum utilization to the optimal solution. Execution of those heuristics with different parameters on two large-sized networks NSFNET and COST-239 reveals that SM-GA outperforms SM-GR while aiming to minimize spectrum and power consumption jointly as well as spectrum utilization only whereas SM-GR performs better than SM-GA while minimizing power consumption only.

在本文中，我们提出了一种新颖的集成方法来设计可生存的离线弹性光网络（EON），以最大程度地降低能耗和频谱需求。通过应用基于多路径的路由和频谱分配（RSA），可以确保单链路故障恢复。最小化频谱利用率的RSA问题是NP-Hard问题。针对可生存的多路径RSA（SM-ILP）制定了整数线性规划模型，以获取小型网络的最佳解决方案。随着SM-ILP解决方案随着网络规模的增长而变得棘手，还开发了两种基于贪婪算法（SM-GR）和遗传算法（SM-GA）的启发式算法，以获得针对大型网络的近乎最优的解决方案。已经进行了模拟研究来评估启发式算法的性能，并将结果与​​通过解决小问题规模的ILP公式所获得的结果进行比较。可以发现，对于考虑的仿真设置，拟议的启发式方法提供的解决方案的能耗变化最大为6.77％，频谱利用率的变化最大为3.71％。在两个大型网络NSFNET和COST-239上执行具有不同参数的试探法表明，SM-GA的性能优于SM-GR，同时旨在最大程度地降低频谱和功耗，并且仅降低频谱利用率，而SM-GR的性能优于SM-GR -GA，同时仅将功耗降至最低。71％的频谱利用率可提供最佳解决方案。在两个大型网络NSFNET和COST-239上执行具有不同参数的试探法表明，SM-GA的性能优于SM-GR，同时旨在最大程度地降低频谱和功耗，并且仅降低频谱利用率，而SM-GR的性能优于SM-GR -GA，同时仅将功耗降至最低。71％的频谱利用率可提供最佳解决方案。在两个大型网络NSFNET和COST-239上执行具有不同参数的试探法表明，SM-GA的性能优于SM-GR，同时旨在最大程度地降低频谱和功耗，并且仅降低频谱利用率，而SM-GR的性能优于SM-GR -GA，同时仅将功耗降至最低。