Applying Fuzzy-Based Inductive Reasoning to Analyze Qualitatively
the Dynamic Behavior of an Ecological System
Abstract
In the past decade, a variety of methodologies for representing and evaluating
knowledge in qualitative terms have been developed, particularly within the
field of Artificial Intelligence. Qualitative reasoning methodologies represent
a valid alternative to quantitative modeling approaches when dealing with
systems for which only imprecise or incomplete knowledge is available, as this
is often the case with ecological systems. As most of the proposed qualitative
reasoning methodologies have not yet been applied to anything but toy examples,
it remains a challenge to apply those methodologies to real-world applications.
In Biosphere 2, a closed ecological system, the level of O2 has dropped and the
CO2 level has risen continuously during its first continuous closure in the
years 1991 to 1993. The mechanisms of atmospheric carbon cycles have been
studied in various research efforts in the past, and general rules describing
them have been formulated. However, the specific situation within Biosphere 2,
a closed ecosystem, might challenge the validity of these rules. The structure
of the carbon cycle in Biosphere 2 is not yet well understood, although abundant
data exist on some of its important fluxes and pools. Whereas deductive,
quantitative as well as qualitative, megies need knowledge about the structure
of the system under study to derive models capturing the behavioral patterns
of the system, the fuzzy-based inductive reasoning methodology FIR learns
inductively the rules of behavior of a system by analyzing input/output data
alone.
Interested in reading the
full paper?
(18 pages, 112,177 bytes, pdf)
Homepage
Last modified: June 15, 2005 -- © François Cellier