Routing and spectrum assignment: A metaheuristic for hybrid ordering selection in elastic optical networks

Abstract

Routing and spectrum assignment (RSA) problem, in elastic optical networks, searches for a route and a sufficient spectrum range to fit each network request. Solving these two problems simultaneously is a complex work. Therefore, they are usually assumed separately by either solving the routing firstly and then the spectrum assignment for this selected route or prioritizing some spectral ranges and then searching for a route that meets such spectrum requirements. The former strategy prioritizes the reduction on the network occupation by using short routes first, while the latter emphasizes load distribution and a more efficient spectrum organization. According to some aspects, such as network topology, number of alternative routes, traffic intensity and number of available slots, emphasizing either set of benefits results in superior advantage. This paper presents a comprehensive study about RSA ordering strategies and proposes a hybrid resource-allocation ordering approach to decide between either strategy. RSA ordering means performing spectrum assignment in function of routing or performing routing in function of some candidate spectrum ranges pre-computed by some SA algorithm criteria An efficient metaheuristic based on genetic algorithm is proposed to implement this strategy, named hybrid routing and spectrum assignment (HRSA). The proposed algorithm returns the most suitable RSA ordering choice for each source–destination pair, which provides near-lowest path-request blocking probability. Simulation results show that the RSA ordering returned by the proposed HRSA reduces the path-request blocking probability up to 35% in comparison to the classical RSA orderings.