Information Technology Reference
In-Depth Information
distributions, constraint decision de-fuzzifi cation, center of area, center
of gravity, fuzzy clustering de-fuzzifi cation, fuzzy mean, generalized level
set de-fuzzifi cation, infl uence value, last or middle of maximum, mean of
maxima, quality method, random choice of maximum, semi-linear de-
fuzzifi cation, weighted fuzzy mean, etc.
Testing of the developed fuzzy model is usually done by predicting the
outputs for input data that were not used for the model construction.
Most software tools, which are available for the application of fuzzy
logic, require the defi nition of fuzzy sets and rules by the user, whereas
more advanced tools, which would automatically develop the rule base
(based on inputs and outputs), are still not widely available. This is the
main disadvantage for the application of fuzzy logic to real-life problems.
It is therefore advisable to use fuzzy logic in combination with other
adaptive systems, such as ANNs to overcome some of the issues
associated. Neuro-fuzzy logic is a hybrid technology that combines the
adaptive learning capabilities of ANNs with the generality of
representation from fuzzy logic (Shao et al., 2007a).
5.2.3 Examples
Fuzzy logic can be applied in both optimization of a pharmaceutical
product (or its certain property) or in the process control. Application of
fuzzy logic in formulation development was described previously (Rowe
and Woolgar, 1999; Rowe and Colbourn, 2000; Shao et al., 2006).
Adaptive fuzzy partitioning was used for prediction of oral
bioavailability (Pintore et al., 2003). Fuzzy logic concepts were also
applied for investigation of hit molecules and molecular
de novo
design
(Klenner et al., 2010). Fuzzy systems have been used to control mechanical
drug delivery devices, in surgical and intensive care settings (Oshita et al.,
1994; Tsutsui and Arita, 1994; Zbinden et al., 1995). A control system
was developed, using fuzzy logic and neural networks, to adjust
intravenous insulin doses in critically ill diabetic patients (Dazzi et al.,
2001). A fuzzy controler was also optimized by a GA, in order to regulate
blood glucose levels in type 1 diabetes (Fereydouneyan et al., 2011).
An adaptive neuro-fuzzy inference system was developed for
optimization of the emulsifi er concentration in the formulation of an
oil-water (o/w) emulsion (Kumar et al., 2010).
Fuzzy logic was used to control granulation processes in agitation fl uid-
bed granulators (Watano et al., 1996). The control of granule growth in
fl uidized bed granulation was carried out using a newly developed image
Search WWH ::
Custom Search