GECCO 2020 Industrial Challenge: Call for Participation

Optimizing Expensive Computational Fluid Dynamics Simulations

GECCO 2020 Logo GECCO 2020 Logo (Bild:

July 8-12, Cancun, Mexico
Sponsored by ACM SIGEVO
Largest Conference in the Field of Genetic and Evolutionary Computation

The goal of the GECCO 2020 Industrial Challenge is to develop an optimizer for this year's industrial application in a programming language of your choice. The optimizer has to efficiently use every objective function evaluation as each evaluation involves a computationally intensive fluid dynamics simulation. As usual, the task from this industrial challenge stems directly from ongoing cooperation with industry.
The challenge is split into two tracks. In the first track, the amount of allowed objective function evaluations is only limited by the computing power you are willing to invest on your own machine. In the second track, only 100 evaluations are allowed per optimization run. In both cases, the algorithm with the best-found objective function value wins.

Like last year, we are able to provide the opportunity for all participants to submit 2-page algorithm descriptions for the GECCO Companion. Thus, it is now possible to create publications in a similar procedure to the Late Breaking Abstracts (LBAs) directly through competition participation!
Highlights of the GECCO 2020 Industrial Challenge include:

  • Interesting Problem Domain: Reduction of emissions and environmental pollution is more important than ever before. Efficient filters help reduce global CO2 emissions.
  • Real-world Problems: Test your algorithms and methods, directly on a real industry problem.
  • Easy Access: Easily Participate through our online platform, no installations required.
  • Fair Submission Assessment: Winners are determined automatically through our online portal, fully objectively, only based on the final result quality.
  • Publication Possibilities: We are able to accept 2-page submissions for the GECCO Companion, thus publications are possible directly through competition participation.

About the Optimization Problem
The Electrostatic Precipitator (ESP) Problem stems from a corporation with Steinmüller Babcock Environment GmbH. The ESP Problem is a 49-dimensional discrete-valued problem. A gas distribution system inside the ESP has to be configured with different types of metal plates, which are expressed in an integer configuration vector. There are eight configurable options for each slot of the gas distribution systems (integers 0-7). Each objective function evaluation involves the automatic setup of a new Computational Fluid Dynamics (CFD) simulation, running that given simulation, and measuring the final obtained gas distribution in the precipitator.

Resource Package 2020

Important Dates and Contacts

  • Software and Data Availability: Online
  • 2-Page Algorithm Description Submission Deadline: 2020-04-03 23:59
  • Final Submission Deadline without Publication Possibility: 2020-06-30 23:59
  • Organizers: F. Rehbach, M. Rebolledo, T. Bartz-Beielstein (TH Köln)
  • Contact: