Combining heuristic and exact approaches for solving the routing and spectrum assignment problem

Efficient algorithms for solving the routing and spectrum assignment (RSA) problem in flex-grid optical networks are of paramount importance to leverage network efficiency. Exact approaches based on integer linear programming (ILP) models are known to be computationally hard and hence, difficult to solve for realistic models while heuristics have inherent downsize of local optimal trapping. In addressing those issues, this paper firstly presents a novel and efficient heuristic based on the genetic algorithm (GA) to provide (near-) optimal solutions to the RSA problem. The effectiveness of the GA-based algorithm is benchmarked against typically used reference heuristics for a set of network scenarios and it is shown that in addition to the better performance than conventional heuristics in the literature, the proposed GA technique indeed could achieve optimal solutions in most cases while the running time is still manageable. Secondly, we propose to make use of efficient solutions from the heuristic approaches to set the upper-bound and warm-start point for ILP formulation. This combination has been shown very productive in enhancing running time performance and increasing the capability of achieving an optimal solution and/or improving the solution quality in a given time span compared with traditional ways for solving ILP models.