Beruflich Dokumente
Kultur Dokumente
Why Fuzzy?
Not all of natural phenomenon can be described using
Crisp. Ex:
Crisp VS Fuzzy
Crisp (yes/no) (a) VS Fuzzy (multi-value) (b)
Crisp VS Fuzzy
Crisp VS Fuzzy
Height 184, in Crisp
it is Tall
Height 184, in Fuzzy
it is Average with
correctness 0.1 and
it is Tall with
correctness 0.4.
Fuzzy Inference
Fuzzy inference can be defined as a process of
mapping from a given input to an output.
There are 2 types of inference system, Mamdanistyle and Sugeno-style
Mamdani-style Inference
Mamdani-style inference
Performed in four steps: fuzzification of the input
variables, rule evaluation, aggregation of the rule
outputs, and finally defuzzification.
Example: A project management case will be
analysed using Mamdani-style inference.
Input: Project funding (PF), project staffing (PS)
Output: Project risk (PR)
PF
PS
Fuzzy (Mamdani-style)
PR
Mamdani-style inference
The fuzzy rules:
Rule: 1
Rule: 1
IF x is A3
IF project_funding is adequate
OR y is B1
OR project_staffing is small
THEN z is C1
THEN risk is low
Rule: 2
Rule: 2
IF x is A2
IF project_funding is marginal
AND y is B2 AND project_staffing is large
THEN z is C2
THEN risk is normal
Rule: 3
Rule: 3
IF x is A1
IF project_funding is inadequate
THEN z is C3THEN risk is high
Mamdani-style inference
Step 1: Fuzzification conducted by providing
fuzzy set for every input and output variables.
Mamdani-style inference
Step 2: Rule evaluation Map the input to every rule and
find the membership value.
IF project_funding is adequate (pf = 35% adequate,
u = 0)
OR project_staffing is small (ps = 60% small, u =
0.1)
THEN risk is low (MAX(0; 0.1) = 0.1) MAX is used
because conjunction of the antecedent is OR.
IF project_funding is marginal (pf = 35% marginal, u
= 0.2)
AND project_staffing is large (ps = 60% large , u =
0.7)
THEN risk is normal (MIN(0.2; 0.7) = 0.2) MIN is used
because conjunction of the antecedent is AND.
Mamdani-style inference
Step 3: Aggregation of the rule output.
Mamdani-style inference
Step 4: Defuzzification Centre of Gravity is the
commonly used method for defuzzification.
Bellow is the function of COG for continuous and
discrete system.
Sugeno-style Inference
Introduction
Mamdani-style inference requires us to find the
centroid of a two-dimensional shape by
integrating across a continuously varying function
not computationally efficient.
Sugeno-style fuzzy inference is very similar to the
Mamdani method.
Sugeno changed only a rule consequent. Instead
of a fuzzy set, Sugeno used a mathematical
function of the input variable.
Format of Sugeno-style
IF x is A
AND y is B
THEN z is f (x, y) the consequent part is
mathematical function of input variables (x and y).
Example:
IF project_funding is High
AND project_staffing is Large
THEN project_risk = 0.7x + 0.85y. (x = crisp input of
project_funding, y = crisp input of project_staffing)
Sugeno-style flow
Sugeno-style defuzzification
Rule: 1
Rule: 1
IF x is A3
IF project_funding is adequate
OR y is B1
OR project_staffing is small
THEN z is C1
THEN risk is 20 (%)
Rule: 2
Rule: 2
IF x is A2
IF project_funding is marginal
AND y is B2 AND project_staffing is large
THEN z is C2
THEN risk is 50 (%)
Rule: 3
Rule: 3
IF x is A1
IF project_funding is inadequate
THEN z is C3THEN risk is 80 (%)
What is the risk if project_funding = 35% and project_staffing = 60% ?
THANK YOU !