Sunteți pe pagina 1din 5

Introduction to Fuzzy-Rule Based

computation
• What is “FUZZY”?? – Antonym of precise/crisp.

• What is “CRISP LOGIC”?? – crisp(precise) inputs -> crisp(precise) outputs

• Why “FUZZY LOGIC”?? – situations may arise where we may not be getting crisp inputs always i.e. We can’t
make crisp outputs for all possible inputs.

• “LINGUISTIC VARIABLE” is an important concept in fuzzy logic.

• In 1965, “Zadeh” proposed the concept of linguistic variables for modelling human way of
thinking(balanced/rational way).
LINGUISTIC VARIABLE AND MEMBERSHIP FUNCTION:

• Linguistic variable – In simple words , it is a variable whose values are in natural language.
• For example, “speed” is a linguistic variable, which can take values as “slow”, “fast”, “very fast”
and so on….
• Each linguistic variable have it’s own belonging function.
• Each is “FUZZY INPUT” is associated with a “degree of membership/belonging” i.e. each value of
linguistic variable has a curve to explain belongingness of the fuzzy input. The membership
function can be of many types triangular, trapezoidal, Gaussian(bell shaped).
• Example: linguistic variable-Age with values={young, middle, old}
The entire fuzzy-rule based computation can be divided into three
Components:

1. FUZZIFICATION:
• It is the process of finding the membership values of a particular fuzzy input with all possible values of
linguistic variables i.e. dividing the input into various parts linguistic variables .

fuzzy input = σ(𝑚𝑒𝑚𝑏𝑒𝑟𝑠ℎ𝑖𝑝 𝑣𝑎𝑙𝑢𝑒𝑠 ∗ 𝑣𝑎𝑙𝑢𝑒𝑠 𝑜𝑓 𝑙𝑖𝑛𝑔𝑢𝑖𝑠𝑡𝑖𝑐 𝑣𝑎𝑟𝑖𝑎𝑏𝑙𝑒)

2. INFERENCE:
• In simple words, it is just application of the rule to all the values the linguistic variable.
• These rules will provide the outputs for all the values of linguistic variables (discrete crisp values).
3. DEFUZZIFICATION:
• This is where the output is computed….
• The outputs of each linguistic variables are multiplied with their corresponding membership values to
produce the final output.

Output = σ(𝑚𝑒𝑚𝑏𝑒𝑟𝑠ℎ𝑖𝑝 𝑣𝑎𝑙𝑢𝑒𝑠 ∗ 𝑜𝑢𝑡𝑝𝑢𝑡𝑠 𝑓𝑜𝑟 𝑒𝑎𝑐ℎ 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑣𝑎𝑟. )

Block diagram:

input output
It’s time to get fuzzyy!!
Let us implement this in making a robot to boil egg precisely.

• As we know that eggs are of various sizes and hence require different boiling times. Let us assume that linguistic variable
be the “size of the egg” and its values be “small(20g)” and “large(70g)”.

• FUZZIFICATION:
It involves finding the membership value of the input egg size with “small” and “large”.
let us assume that membership values be ‘X’{For Small} and ‘Y’{For Large}

• INFERENCE:
Finding the boiling time for small and large egg from the Rulebook. Let it be 3 min and 6 min respectively.
• DEFUZZIFICATION:
Let the boiling egg of the given egg be ‘t’
t= X*3 + Y*6

S-ar putea să vă placă și