The 25th Workshop Computational Intelligence 2015, an annual conference held in Nov'2015 by the Computational Intelligence (CI) Chapter of VDI-GMA (Gesellschaft für Mess- und Automatisierungstechnik) in Dortmund, has attributed the Young Author Award to Samineh Bagheri, PhD, scientific member of the research group of Professor Konen at Campus Gummersbach. Congratulations!

She got this award for the contribution

SACOBRA: Self-Adjusting Constrained Black-Box Optimization with RBF

which she wrote as main author, conjointly with Wolfgang Konen, Dr. Patrick Koch, former member of Wolfgang Konen's research group, Prof. Dr. Thomas Bäck, Leiden University + divis intelligent solutions GmbH, and further colleagues from divis intelligent solutions GmbH. Samineh presented this work in a well-received talk.

The above snapshot shows the jurors and the Young Author Award winners 2015, from left to right: juror Frank Hoffmann, professor at TU Dortmund, M. Sc. Benjamin Schott, Karlsruher Institut für Technologie (KIT), M. Sc. Samineh Bagheri, TH Köln (Cologne University of Applied Sciences), M. Sc. Sarah Uttendorf, IPH Hannover, juror Andreas Kroll, professor at University Kassel.

The Young Author Award is an annual price since the year 2000 and it is the for second time attributed to a member from Cologne University of Applied Sciences (after Patrick Kochs Young Author Award last year).

 

The research leading to this contribution took place in the context of the ZIM research project MONREP. The theme of this work is in the context of optimization problems in high dimensions with many constraints and how they can be solved efficiently, i. e. with a low number of function evaluations.  This has a very realistic application background: In automotive industry it is desirable to minimize the weight of a car, but at the same time the car body has to meet all safety requirements. Each car design has to be tested carefully in very time-consuming simulations. This is the reason why only a low number of simulations (= function evaluations) is possible. The car designs are of course company secrets, but an equivalent benchmark with name MOPTA08 was released by the automotive industry and is one of the topics of our research. More details can be read in the preprint of the CI-workshop contribution.

MONREP was initiated by Prof. Thomas Bäck (divis GmbH) and Prof. Wolfgang Konen (Cologne University of Applied Sciences) and has a duration of 2.5 years (2014-2016). Samineh Bagheri, M.Sc., works for MONREP as a PhD-researcher in part-time, the other part she works as teaching assistant.

More details on MONREP in Pressemitteilung MONREP (PDF) (sorry, in German only!) and on the MONREP project page in CIOP.

poster

The research group supervised by Prof. Wolfgang Konen from TH Köln has a collaboration with the Natural Computing group (head: Prof. Thomas H.W. Bäck) of the LIACS Computer Science department at Leiden University. On October,16th, 2015, a seminar took place in Leiden, Netherlands, in which both groups presented their recent research works.

The main topics of this seminar were:

  • Multi-objective optimization
  • Surrogate assisted global optimization
  • Modeling techniques (Radial Basis Functions, Kriging)

Our group gave three talks to present our recent progress in developing a software for constrained optimization problems, we also discussed the existing challenges for solving such a problem. All three talks are listed as follows:

  •  S. Bagheri, W. Konen: A new repair method for constrained optimization.
  •  W. Konen, S. Bagheri: Common pitfalls in RBF modeling.
  •  S. Bagheri, W. Konen: Solving the G-problems in less than 500 iterations: Better understanding of surrogate modeling improves efficient constrained optimization.

The other three talks given by Dr. Emmerich (Assistant Professor at Leiden University) , Hao Wang and Bas van Stein (PhD students at Leiden University)  are as follows:

  • M. Emmerich: From Data to Decisions: Recent Advances and Future Challenges in Multicriteria Optimization and Decision Making
  • H. Wang: Multi-point Efficient Global Optimization using Niching Evolution Strategy
  • B. van Stein: Cluster Kriging

Interesting talks, lively discussions, friendly atmosphere and an active and motivated research team made last Friday very productive. 


madridConference

 

 

 

 

 

 

 

 

This year GECCO took place in Madrid on 11-15 of July. Many interesting tutorials and workshops were presented in this conference. On 14th of July we presented our paper about a new repair method for constrained optimization. You can download the presentation slides from here.

 

SACOBRApackage

 

 

 

 

 

 

 

 

We are glad to inform you that SACOBRA package is now available on CRAN. 

SACOBRA package performs surrogate-assisted optimization by utilizing radial basis function interpolation. SACOBRA solver is a derivative-free optimizer appropriate to address constrained expensive optimization problems.

TD-TicTacToe2

A new technical report on temporal difference (TD) learning for games and "self-play" algorithms for game-agent training is available. This report by Wolfgang Konen features a gentle introduction to TD learning for game play and gives hints for the practioner on the implementation of such algorithms . It shows the references to the most recent applications in this field and discusses in an appendix the more advanced topic of eligibility traces and how and why they work.

This report should be a help for people starting new in the field of TD learning for games and for people who work already in this field but struggle with specific details. It is an updated English translation of an earlier report in German language.

Publications

Konen, Wolfgang (2015): Reinforcement Learning for Board Games: The Temporal Difference Algorithm. Research Center CIOP (Computational Intelligence, Optimization and Data Mining) Cologne University of Applied Sciences, 2015.

PDF English

PDF German