This paper studies the restricted vertex 1-center problem (RV1CP) and restricted absolute 1-center problem (RA1CP) in general undirected graphs with each edge having two weights, cost and delay. First, we devise a simple FPTAS for RV1CP with running time, based on FPTAS proposed by Lorenz and Raz (Oper. Res. Lett. 28(1999), 213–219) for computing end-to-end restricted shortest path (RSP). During the computation of the FPTAS for RV1CP, we derive a RSP distance matrix. Next, we discuss RA1CP in such graphs where the delay is a separable (e.g., linear) function of the cost on edge. We investigate an important property that the FPTAS for RV1CP can find a (1+ϵ)-approximation of RA1CP when the RSP distance matrix has a saddle point. In addition, we show that it is harder to find an approximation of RA1CP when the matrix has no saddle point. This paper develops a scaling algorithm with at most (Formula Presented.) running time where K is a step-size parameter and η is a given positive number, to find a (1+η)-approximation of RA1CP.
CITATION STYLE
Ding, W., & Qiu, K. (2015). Approximating the restricted 1-center in graphs. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9486, pp. 647–659). Springer Verlag. https://doi.org/10.1007/978-3-319-26626-8_47
Mendeley helps you to discover research relevant for your work.