Neural Networks and Evolutionary Algorithms for the Prediction of Thermodynamic Properties for Chemical Engineering

  • Martin Mandischer
  • Hannes Geyer
  • Peter Ulbig
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 1585)


In this paper we report results for the prediction of thermo-dynamic properties based on neural networks, evolutionary algorithms and a combination of them. We compare backpropagation trained networks and evolution strategy trained networks with two physical models. Experimental data for the enthalpy of vaporization were taken from the literature in our investigation. The input information for both neural network and physical models consists of parameters describing the molecular structure of the molecules and the temperature. The results show the good ability of the neural networks to correlate and to predict the thermodynamic property. We also conclude that backpropagation training outperforms evolutionary training as well as simple hybrid training.


Neural Networks Evolution Strategies Hybrid-Learning Chemical Engineering 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    T. Bäck. Evolutionary Algorithms in Theory and Practice. Oxford Univ. Press, New York, 1996.zbMATHGoogle Scholar
  2. 2.
    L. M. Egolf and P. Jurs. Prediction of boiling points of organic heterocyclic compounds using regression and neural network techniques. In J. Chem. Inf. Comput. Sci. 33, pages 616–625. 1993.CrossRefGoogle Scholar
  3. 3.
    H. Geyer, P. Ulbig, and S. Schulz. Encapsulated evolution strategies for the determination of group contribution parameters in order to predict thermodynamic properties. In 5th Int’l. Conf. on Parallel Problem Solving from Nature. Amsterdam, 1998.Google Scholar
  4. 4.
    K. Hornik, M. Stinchcombe, and H. White. Multilayer feedforward networks are universal approximators. Neural Networks, 2:359–366, 1989.CrossRefGoogle Scholar
  5. 5.
    M. Mandischer. Evolving recurrent neural networks with non-binary encoding. In Proc. Second IEEE Int’l Conf. Evolutionary Computation (ICEC’ 95), vol. 2, pages 584–589, Perth, Australia, 1995. IEEE Press, Piscataway NJ.Google Scholar
  6. 6.
    D. E. Rummelhart and J. L. McClelland. PDP: Explorations in the Microstructure of Cognition, volume 1. MIT Press, Cambridge, MA, USA, 1986.Google Scholar
  7. 7.
    H.-P. Schwefel. Evolution and Optimum Seeking. Sixth-Generation Computer Technology. Wiley, New York, 1995.Google Scholar
  8. 8.
    P. Ulbig. Gruppenbeitragsmodelle UNIVAP & EBGCM. Dr.-Ing. Thesis, Univ. of Dortmund, Institute for Thermodynamics, 1996.Google Scholar
  9. 9.
    P. Ulbig, T. Friese, H. Geyer, C. Kracht, and S. Schulz. Prediction of thermodynamic properties for chemical engineering with the aid of Computational Intelligence. In Progress in Connectionist-Based Information Systems. Springer, 1997.Google Scholar
  10. 10.
    W. Wienholt. Minimizing the system error in feedforward neural networks with evolution strategy. In S. Gielen and B. Kappen, editors, Proc. of the Int’l. Conf. on Artificial Neural Networks, pages 490–493, London, 1993. Springer-Verlag.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1999

Authors and Affiliations

  • Martin Mandischer
    • 1
  • Hannes Geyer
    • 2
  • Peter Ulbig
    • 2
  1. 1.Department of Computer Science XIUniversity of DortmundGermany
  2. 2.Institute for ThermodynamicsUniversity of DortmundGermany

Personalised recommendations