Video summary

Logika Fuzzy [1]: Fungsi Keanggotaan

Main summary

Key takeaways

Educational

Main Ideas and Concepts (Fuzzy Logic Overview)

Motivation: Handling Uncertainty

  • Traditional logic is binary (0/1, black/white).
  • Real-world judgments are often uncertain or gradual, such as:
    • “fast vs slow”
    • “tall vs not tall”
  • Fuzzy logic models uncertainty using degrees of truth rather than only true/false.

Core Analogy Examples

  • Speed comparisons
    • Example speeds: motorbike 15, car 50, duck 5
    • Fuzzy logic helps represent terms like fast/slow when boundaries are unclear.
  • Human height
    • Around a threshold (e.g., 170, 140, 171 cm)
    • Shows how “tall” can be ambiguous near cutoff values.

Fuzzy Logic Definition

  • Truth values range from 0 to 1 (not only 0 or 1).
  • Uses linguistic variables (e.g., “cold/hot”, “young/old”, “tall”).
  • Each linguistic variable is represented by membership functions that map inputs to degrees of membership.

Methodology / Structured Learning Points

A. Fuzzy System Lifecycle (High-Level Process)

  1. Fuzzification
    • Convert crisp inputs (exact numeric values) into linguistic variables and their membership degrees.
  2. Inference (Fuzzy Inference)
    • Apply rules from the knowledge base using fuzzy set operations to produce intermediate results (e.g., max/min).
  3. Defuzzification
    • Convert the fuzzy output back into a crisp numeric output.

B. Fuzzy Sets Basics (Building Blocks)

Fuzzy Variables

  • A fuzzy system includes variables such as income, temperature, demand, age, height, speed.

Fuzzy Sets

  • For each variable, fuzzy sets represent conditions like:
    • “cold/hot” (temperature)
    • “young/old” (age)

Attributes of a Fuzzy Set

  1. Linguistic: the word label (e.g., cold/hot, young/old)
  2. Numeric / membership degree mapping: membership strength as values in [0, 1]
  3. Universe of discussion: allowed input range for that variable
    • Example: temperature from -10 to 90°C

Membership Function (μ)

  • A graph/function that outputs the degree of membership for each input.
  • Membership is not restricted to 0 or 1 (e.g., 0.6, 0.7, etc.).

C. Membership Function Shapes (Computation Approach)

The content describes multiple membership function types and emphasizes using their equations depending on which interval the input falls into.

1. Triangular / Linear Segment Style (Increasing and Decreasing)

  • Increasing line
    • Limited by parameters a and b
    • Membership changes linearly between those points.
  • Decreasing line
    • Again limited by a and b
    • Membership linearly falls across that interval.

Key interval logic

  • If input is before the left boundary, membership becomes 0 (or 1, depending on the side of the curve).
  • If input is between boundaries, compute with the appropriate linear equation.
  • If input is after the right boundary, membership becomes 0 (or 1, depending on which side applies).

Example workflow shown

  • Determine where x lies relative to a and b
  • Use the correct piecewise linear formula
  • Example result: for an increasing curve with a = 25, b = 35, membership of x = 32 computed as 0.7

2. Trapezoidal Curve

  • Parameters: a, b, c, d
  • Structure:
    • Rising edge: linear increase from a to b
    • Plateau: membership at 1 between b and c
    • Falling edge: linear decrease from c to d

Computation guidance

  • If x is on the plateau, membership is 1
  • If x is on the rising/falling edge, use the corresponding edge equation
  • If x is outside [a, d], membership becomes 0

Edge-case rule

  • If x equals a boundary such as c, membership can be treated as 1 and should match the visual interpretation.

3. S-shaped (Sigmoid) Curves

  • Used for smooth transitions (gradual membership change).
  • Determine where x sits relative to key points (often via midpoints/interval partitions such as a, b, c).
  • Decide whether x is left or right of the midpoint(s).
  • Use the sigmoid equation for the correct interval.

Example given

  • Sigmoid growth produced a membership result of 0.6.

4. Sigmoid Shrinkage (Opposite Direction)

  • Same smooth-transition concept, but membership decreases.
  • Example given: membership result of 0.3 for shrinking.

5. Bell-Shaped Curve

  • Described as a combination of sigmoid growth and shrinkage.
  • Shape: a “hump” (peak membership in the middle, lower on both sides).
  • Example practice mentioned: compute membership values for points like 45 and 33 using a bell-shaped curve.

6. Concave / “Gaussian-like” or Other Curved Variations

  • Mentions additional bell-shaped options (e.g., concave curve).
  • Suggests they can be learned from independent reference materials.

D. Fuzzy Set Operations (Used in Inference)

The video lists three core operations:

  • Join / OR (Max)
    • Rule: membership = max(a, b)
  • Meet / AND (MIN)
    • Rule: membership = min(a, b)
  • Complement / NOT
    • Rule: membership = 1 − a

These operations are then used inside inference rules.

E. Example of Inference Rule Behavior (Conceptual)

  • If a rule uses OR between two antecedents:
    • Combine via max
  • If a rule uses AND:
    • Combine via min
  • If a rule uses NOT / complement:
    • Use 1 − membership

Applications Mentioned (Why Fuzzy Logic Is Useful)

  • Consumer electronics
    • Air conditioners: adjust comfort temperature even if the number of people changes.
    • Washing machines: sensors detect dirt level/type to choose wash/spin settings automatically.
    • Cameras/lighting systems: adjust light intensity entering the camera.
  • Automotive
    • Fuzzy systems in automatic transmission modes; claimed fuel savings (~12–17%).
  • Engineering
    • Predicting earthquake timing based on research (not just assumptions).
  • Medicine and biology
    • Diagnosing cancer
    • Assisting prosthetics fit
    • Supporting vaccine search for outbreaks/pandemics
  • Environmental science
    • Weather prediction such as rain or hurricane occurrence

Learning Emphasis / Lesson Conveyed

  • Begin with fundamentals: fuzzy variables, fuzzy sets, membership functions.
  • Learn by understanding piecewise interval logic:
    • Identify which region of the membership function the input falls into
    • Apply the equation for that region
  • Membership function equations don’t need to be blindly memorized—understanding the curve and intervals matters most.

Speakers or Sources Featured

  • Presenter/Instructor: Fauzi (the narrator/instructor)
  • Professor: Lutfi (Zadeh) (credited as initiating fuzzy logic concepts)
  • Video/Platform sources: Google (referenced via “OK Google” in one example)

Original video