Routing, modulation level, and spectrum assignment in elastic optical networks. A route-permutation based genetic algorithms

The routing, modulation level, and spectrum allocation (RMLSA) problem is crucial for efficient elastic optical networks. This problem has been approached by optimal-and-non-scalable and sub-optimal-and-scalable solutions. In the second approach, we can distinguish the routing-based and permutation-based meta-heuristics. These approaches explore a sub-set of the RMLSA solutions, and consequently, the calculation of high-quality solutions can be limited.

This work proposes an RMLSA solution that considers the routing and request permutation simultaneously to explore a larger portion of the set of RMLSA solutions than state-of-the-art meta-heuristics. The proposed RMLSA solution is based on a genetic algorithm (GA) whose chromosome structure encodes routing and permutation genes.

Performance analysis of the proposed route-permutation-based GA (RPGA) has been compared to the state-of-the-art based on integer linear programming (ILP), route-based GA (RGA), and permutation-based GA (PGA) in offline and online traffic scenarios. Offline traffic simulations show that RPGA is promising since it obtains similar results to ILP. RGA gets worst as the traffic load increases compared to PGA and RPGA approaches. RGA, PGA, and RPGA achieve the same performance in all dynamic scenarios concerning blocking and entropy measures, given the set of requests is small.