Prof. Dr. Tobias Friedrich


Two papers accepted at GECCO

This years Genetic and Evolutionary Computation Conference (GECCO) will, due to the world-wide coronavirus pandemic, be held online in mid July. Traditionally, this world-leading venue hosts research on nature-inspired algorithms and evolutionary computation. The Algorithm Engineering group contributes two papers to this years edition.

As an intermediate step between the well understood TSP (traveling salesman problem) and the more complicated TTP (traveling thief problem), Bossek et al. study a version of TSP where city weights are fixed and the cost of traveling increases with respect to the weights of cities already visited during a tour.

Doerr and Krejca show that a simple bivariate estimation-of-distribution algorithm (EDA) is capable of efficiently generating a probabilistic model of the search space that is near optimal and implicitly represents an exponential number of optima. This behavior is impossible for population-based evolutionary algorithms or univariate EDAs.

  • The Node Weight Dependent... - Download
    Bossek, Jakob; Casel, Katrin; Kerschke, Pascal; Neumann, Frank The Node Weight Dependent Traveling Salesperson Problem: Approximation Algorithms and Randomized Search HeuristicsGenetic and Evolutionary Computation Conference (GECCO) 2020: 1286–1294
  • Bivariate Estimation-of-D... - Download
    Doerr, Benjamin; Krejca, Martin S. Bivariate Estimation-of-Distribution Algorithms Can Find an Exponential Number of OptimaGenetic and Evolutionary Computation Conference (GECCO) 2020: 796–804