Problem landscapes of various kinds, such as fitness landscapes, violation landscapes, and multiobjective problem landscapes, are increasingly being studied and exploited in search and optimization. They may serve multiple purposes. By using features derived from the landscapes, one can characterize and better understand the problem at hand. Landscapes may also be used to explain the behavior of algorithms and predict their performance. Based on this approach, one can select or configure appropriate algorithms as well.

This special session aims to strengthen research on problem landscape analysis and its exploitation in solving real-world problems. Its purpose is to bring together researchers and practitioners interested in contributing to this challenging area and benefiting from its most recent achievements. Theoretical and empirical studies, as well as examples from practical applications, are all welcome.

Download the CFP in PDF here.

Topics of Interest

Topics of interest include, but are not limited to:

  • Fitness/violation/multiobjective problem landscapes
  • Sampling for problem landscape analysis
  • Landscape feature selection/construction
  • Exploratory landscape analysis
  • Local optima networks
  • Search trajectory networks
  • Problem characterization/understanding
  • Machine learning on landscape data
  • Landscape metrics 
  • Algorithm selection/configuration/construction
  • Algorithm performance prediction
  • Landscape visualizations
  • Applications of problem landscape analysis


  • Bogdan Filipič
    Jožef Stefan Institute, Slovenia
  • Pavel Krömer
    Technical University of Ostrava, Czech Republic