You are on page 1of 24

Fuzzy logic

Linguistic constructs
Fuzzy system
Crisp set
• short - linguistic variable
– X=> 4.5 inch
– Y=>3.5 inch
– Z=>4.0 inch

• Curve varies linearly between 0 and 1


• These rule-based systems can in theory model
any system with arbitrary accuracy, that is,
they work as universal approximators.
• The number of rules increases exponentially
with the dimension of the input space
(number of system variable)
crisp sets
Knowledge systems
• One approach to representing and processing
of knowledge is logic.
• Two-state (bivalent), crisp logic uses only two
quantities,true (T) and false (F) as truth
values.
Find its crisp Truth value?

• “All my statements are lies”

• Ans:
Stages of development of an
intelligent product
• Conception of the idea of the product and discussion
with interested parties for product development.
Obstacles: insufficient knowledge on soft computing,
AI, knowledge-based systems…
• Demonstration of the concept using simplified
examples. obstacles: skepticism of co-workers of
managers, potential sponsors, and users and over
expectation of product remedy: Feasibility
demonstrations by building a small prototype system
to demonstrate that a knowledge system to avert
Stages of development of an
intelligent product
• Development of practical specifications and
designing a product prototype to satisfy the
specifications. obstacle: shortage of technical
personnel . Remedy :The use of flexible, portable,
and commonly available tools and techniques is
desirable
• Integration of the product into an existing
system.- technology transfer – worker fear on NT
- retrain and interact with R&D personnel,
• managers, operators, maintenance personnel,
etc.
Use of fuzzy logic in expert systems
• Experts may be only partially knowledgeable about
the problem domain, or data may not be fully
available, yet decisions are still expected.
• Why to use?:
– The knowledge base of expert systems encapsulates the
knowledge of human experts.
– Human knowledge is often inexact and qualitative, and
fuzzy descriptors (e.g., large, small, fast, poor, fine) are
commonly used in its communication
– Problem description of the user may not be exact
– Knowledge base from experts may not be complete, yet
satisfactory decisions are still expected
– Educated guesses need to be made in some situations
Fuzzy sets
• No crisp boundary , no binary membership characteristics -
partial membership(softness) is feasible
• Eg: temperature( e.g., cold, cool, tepid, warm, hot)=> each
element is a fuzzy descriptor
• Fuzzy Set:
Membership function
• fuzzy set represented by a membership
function[μA(x)] – defines grade (degree) of
membership (possibility (possibility function) )
maps each element[x] of the universe[X] to value
[0, 1].
• μA(x): X → [0, 1]
• Or fuzzy set A is a set of ordered pairs.
• A = {(x, μA(x)); x ∈X, μA(x) ∈[0, 1]}
• fuzziness describes the ambiguity of an event and
randomness describes the uncertainty in
occurrence of event
• A fuzzy set contains an object x to degree a(x), that is, a(x) =
Degree(x E A)
• The fuzzy set A can be expressed as
• A = {(x, a(x))}, x E X;
• imposes an elastic constraint of the possible values of
elements x EX, called the possibility distribution.
• fuzziness describes the ambiguity of an event and
randomness describes the uncertainty in occurrence of
event

• Note: E= belongs symbol


• a=1, c=0,b=vague non-crisp (or fuzzy)
membership
• Closer to μA(x) is to 1
– x is near A
• Symbolic representation:
X= All students in class
Probability vs Fuzzy
Fuzzy set vs. Crisp set

You might also like