Endre Csóka

  1. Local algorithms with public randomisation on sparse graphs.

    Authors: Endre Csóka
    Subjects: Combinatorics
    Abstract

    Consider the problem when we want to construct some structure on a bounded
    degree graph, e.g. an almost maximum matching, and we want to decide about each
    edge depending only on its constant radius neighbourhood. We show that the
    information about the local statistics of the graph does not help here at all.
    Namely, if there exists a local algorithm which can use any local statistics
    about the graph, and produces a good approximation for a parameter, then there
    exists an approximation algorithm not using any statistics.

  2. Maximum flow is approximable by deterministic constant-time algorithm in sparse networks.

    Authors: Endre Csóka
    Subjects: Data Structures and Algorithms
    Abstract

    We show a deterministic constant-time parallel algorithm for finding an
    almost maximum flow in multisource-multitarget networks with bounded degrees
    and bounded edge capacities. As a consequence, we show that the value of the
    maximum flow over the number of nodes is a testable parameter on these
    networks.

Syndicate content