Network design with node connectivity constraints

Abstract

The Generalized Steiner Problem with Node-Connectivity constraints (GSPNC) consists of determining of a minimum cost subnetwork of a given network where some pairs of nodes are required to satisfy node-connectivity requirements. The GSPNC has applications to the design of low-cost communications networks which can survive failures in the servers as well as in the connection lines. The GSPNC is known to be NP-Complete. In this paper, we introduce an algorithm based on GRASP (Greedy Randomized Adaptive Search Procedure), an effective combinatorial optimization metaheuristic. Experimental results are obtained over a set of problem instances with different characteristics and connectivity requirements, obtaining in all these cases optimal or near-optimal results.