Sequential parameter optimization can be described as a tuning algorithm with the following properties:spotOfficial

  • (i) Use the available budget (e.g., simulator runs, number of function evaluations) sequentially, i.e., use information from the exploration of the search space to guide the search by building one or several meta models. Choose new design points based on predictions from the meta model(s). Rene the meta model(s)) stepwise to improve knowledge about the search space.
  • (ii) Try to cope with noise by improving condence. Guarantee comparable confidence for search points.
  • (iii) Collect information to learn from this tuning process, e.g., apply explorative data analysis.
  • (iv) Provide mechanisms both for interactive and automated tuning.

SPOT was developed over the last years by Thomas Bartz-Beielstein, Christian Lasarczyk, and Mike Preuss. The main purpose of SPOT is to determine improved parameter settings for optimization algorithms to analyze and understand their performance.

SPOT was successfully applied to numerous optimization algorithms, especially in the field of evolutionary computation, i.e., evolution strategies, particle swarm optimization, algorithmic chemistries etc. in the following domains:

  • machine engineering: design of mold temperature control
  • aerospace industry: airfoil design optimization
  • simulation and optimization: elevator group control
  • technical thermodynamics: non sharp separation
  • economy: agri-environmental policy-switchings
  • algorithm engineering: graph drawing
  • statistics: selection under uncertainty (optimal computational budget allocation) for PSO
  • evolution strategies: threshold selection and step-size adaptation
  • genetic chromodynamics
  • computational intelligence: algorithmic chemistry
  • particle swarm optimization: analysis and application
  • numerics: comparison and analysis of classical and modern optimization algorithms
  • vehicle routing and door-assignment problems
  • bioinformatics
  • storm water management
  • differential and integral equations
  • time series analysis

An R version of this toolbox for interactive and automatic optimization of algorithms can be downloaded from CRAN.


Thomas Bartz-Beielstein, Christian Lasarczyk, and Mike Preuss