We address a variant of the classical Steiner tree problem defined over undirected graphs. The objective is to determine the Steiner tree rooted at a source node with minimum cost and such that the number of edges is less than or equal to a given threshold. The link constrained Steiner tree problem (LCST P) belongs to the NP-hard class. We formulate a Lagrangian relaxation for the LCST P in order to determine valid bounds on the optimal solution. To solve the Lagrangian dual, we develop a dual ascent heuristic based on updating one multiplier at time. The proposed heuristic relies on the execution of some sub-gradient iterations whenever the multiplier update procedure is unable to generate a significant increase of the Lagrangian dual objective. We calculate an upper bound on the LCST P by adjusting the infeasibility of the solution obtained at each iteration. The solution strategy is tested on instances inspired by the scientific literature.
An algorithm to find the link constrained steiner tree in undirected graphs
Di Puglia Pugliese Luigi;
2016
Abstract
We address a variant of the classical Steiner tree problem defined over undirected graphs. The objective is to determine the Steiner tree rooted at a source node with minimum cost and such that the number of edges is less than or equal to a given threshold. The link constrained Steiner tree problem (LCST P) belongs to the NP-hard class. We formulate a Lagrangian relaxation for the LCST P in order to determine valid bounds on the optimal solution. To solve the Lagrangian dual, we develop a dual ascent heuristic based on updating one multiplier at time. The proposed heuristic relies on the execution of some sub-gradient iterations whenever the multiplier update procedure is unable to generate a significant increase of the Lagrangian dual objective. We calculate an upper bound on the LCST P by adjusting the infeasibility of the solution obtained at each iteration. The solution strategy is tested on instances inspired by the scientific literature.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.