Evaluation of self-organizing systems using quantitative measures


Evaluation of self-organizing systems using quantitative measures

Richard Holzer and Hermann de Meer -University of Passau


For analyzing properties of self-organizing systems, mathematical models
of the systems are considered desirable. Such models can be used for suggesting
improvements to a real system under investigation or for gaining confidence
in the purposeful behavior of self-organizing systems.
Global properties of interest that can be related to purposefulness, like
emergence or target orientation, may also be identified on the micro-level,
such that quantitative measures do provide a link from micro-level
to macro-level modeling.  Due to the high computational complexity of practical
systems, however, it is usually prohibitive to derive the exact values of the
measures. Therefore, approximations have been developed, which are
presented and discussed in this talk. Limits and potentials of the
new concepts are illustrated by means of some example applications.


About the Speaker: