Dirk Sudholt

  1. Adaptive Population Models for Offspring Populations and Parallel Evolutionary Algorithms.

    Authors: Jörg Lässig, Dirk Sudholt
    Subjects: Computational Complexity
    Abstract

    A central question when parallelizing evolutionary algorithms is the choice
    of the number of parallel instances. In practice optimal parameter settings are
    often hard to find due to limited information about the optimization problem
    under consideration. We present two adaptive schemes for dynamically choosing
    the number of instances in each generation. These schemes work in a black-box
    setting where no knowledge on the function at hand is available.

Syndicate content