Approximation Algorithms for Disjoint st-Paths with Minimum Activation Cost

Abstract

In network activation problems we are given a directed or undirected graph G = (V,E) with a family {f _uv (x _u ,x _v ): (u,v) ∈ E} of monotone non-decreasing activation functions from D ² to {0,1}, where D is a constant-size domain. The goal is to find activation values x _v for all v ∈ V of minimum total cost ∑  _v ∈ V x _v such that the activated set of edges satisfies some connectivity requirements. Network activation problems generalize several problems studied in the network literature such as power optimization problems. We devise an approximation algorithm for the fundamental problem of finding the Minimum Activation Cost Pair of Node-Disjoint st-Paths (MA2NDP). The algorithm achieves approximation ratio 1.5 for both directed and undirected graphs. We show that a ρ-approximation algorithm for MA2NDP with fixed activation values for s and t yields a ρ-approximation algorithm for the Minimum Activation Cost Pair of Edge-Disjoint st-Paths (MA2EDP) problem. We also study the MA2NDP and MA2EDP problems for the special case |D| = 2.