Solving the Generalized Steiner Problem in edge-survivable networks

Abstract

The Generalized Steiner Problem with Edge-Connectivity constraints (GSP-EC) consists of computing the minimal cost subnetwork of a given feasible network where some pairs of nodes must satisfy edge-connectivity requirements. It can be applied in the design of communications networks where connection lines can fail and is known to be an NP-Complete problem. In this paper we introduce an algorithm based on GRASP (Greedy Randomized Adaptive Search Procedure), a combinatorial optimization metaheuristic that has proven to be very effective for such problems. Promising results are obtained when testing the algorithm over a set of heterogeneous network topologies and connectivity requirements; in all cases with known optimal cost, optimal or near-optimal solutions are found.