Advertisement

Neural Networks in Missing Data Analysis, Model Identification and Non Linear Control

  • S. Salini
  • T. Minerva
  • A. Zirilli
  • A. Tiano
  • F. Pizzocchero
Conference paper
Part of the Advances in Soft Computing book series (AINSC, volume 18)

Abstract

This paper is concerned with the problem of designing a stochastic control system for a wastewater treatment plant. Such a problem brings a sort of complexity making the task difficult to solve owing to the presence of missing data, strong interactions among the pollution variables, feedback and non-linear effects. To undertake this problem a hybrid methodology is proposed. Firstly, we addressed the problem of missing data imputation and model identification by considering classes of neural networks as potential approximated models. The optimal neural predictive model selection was obtained within an evolutionary approach (EvoNeural Model). A global search evolutionary technique is then used to establish and tune the control system related to the selected EvoNeural Model (EvoNeural Control Model). The resulting EvoNeural Control Model is then applied to a wastewater treatment plant in order to obtain an experimental validation of the proposed methodology.

Keywords

Neural Network Genetic Algorithm Artificial Neural Network Wastewater Treatment Plant Carbon Oxide 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. BOX GEORGO E. P. and JENKINS M., (1994), Time series analyis: forecasting and control, Englewood Cliffs, N. J.: Prentice HallGoogle Scholar
  2. CYBENKO, G.,(1989), Approximation by superpositions of sigmoidal function, Mathematical of control, Signals, and sistems, 2, 303 – 314Google Scholar
  3. GOLDBERG, D.E. (1989), Genetic Algorithms in search, optimisation and machine learning, Addison-Wesley Publishing CoGoogle Scholar
  4. HOLLAND, J.H. (1975), Adaptation in Natural and Artificial Systems, University of Michigan Press, Ann ArborGoogle Scholar
  5. LJUNG, L. (1987). System Identification: Theory for the User, Prentice-Hall, LondonMATHGoogle Scholar
  6. MINERVA, T., PATERLINI, S. and POLI, I, (1999), Hybrid algorithms for time series analysis: applications to financial data, accepted for pubblication by Economics and Complexity, 3-4, 57 – 77Google Scholar
  7. MITCHELL, M. (1996), An introduction to Genetic algorithms, MIT Press, Cambidge, MAGoogle Scholar
  8. NARENDRA K.S. (1992), Adaptive control of dynamical systems using neural networks, Handbook of Intelligent Control, New York, Van Nostrand Reionhold, pp. 141 – 183.Google Scholar
  9. NARENDRA K.S. and BALAKRIAHNAN S. (1997), Adaptive control using multiple models, IEEE Transaction on Automatic Control (AC-42),pp.171-187.Google Scholar
  10. SALINI S., MINERVA T., PECORARO A., PIZZOCCHERO F., AND TIANO A. (2001), Comparison between neural net models and stochastic models on identification of a wastewater treatment plant, AMSDA2001.Google Scholar
  11. ZIRILLI A., ROBERTS G.N., TIANO A., SUTTON R. (2000), Adaptive steering of a containership based on neural networks, Int. J. Adaptive Control and Signal Processing, (14), pp. 849 – 873.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2003

Authors and Affiliations

  • S. Salini
    • 1
  • T. Minerva
    • 2
  • A. Zirilli
    • 3
  • A. Tiano
    • 3
  • F. Pizzocchero
    • 3
  1. 1.Istituto di StatisticaUniversità Cattolica del Sacro Cuore di MilanoItaly
  2. 2.Dipartimento di Scienze CognitiveSociali e Quantitative Università di Modena e Reggio EmiliaItaly
  3. 3.Dipartimento di Informatica e SistemisticaUniversità di PaviaItaly

Personalised recommendations