12.1 Many-Valued and Fuzzy Logics

Many-valued logics expand on classical logic's true/false dichotomy, introducing additional truth values to handle vagueness and uncertainty. This approach addresses real-world complexities like subjective statements and borderline cases that classical logic struggles with.

Fuzzy logic takes this further, allowing infinite truth values between 0 and 1. It uses fuzzy sets, linguistic variables, and fuzzy operators to represent degrees of truth and make decisions based on natural language rules.

Many-Valued Logics

Limitations of classical logic

• Classical two-valued logic assigns propositions as either true or false
  • Principle of bivalence states every proposition must be either true or false with no intermediate values
  • Struggles to handle vagueness, uncertainty, and propositions with degrees of truth (tall, hot, nice weather)
    • Unable to represent propositions that are partially true or have borderline cases
    • Ambiguity and context-dependent statements are difficult to represent (bald, heavy, "The book was interesting")
• Real-world examples highlight limitations of classical logic
  • "The weather is nice today" is subjective and depends on individual preferences and context
  • "John is tall" is vague since height is a continuum and "tall" is not precisely defined (6 feet, 2 meters)
• Alternative logics like many-valued and fuzzy logic developed to address limitations of classical logic

Principles of many-valued logics

• Many-valued logics introduce additional truth values beyond true and false
• Three-valued logic, such as Łukasiewicz's logic, includes a third value
  • Truth values are true, false, and unknown or indeterminate
  • Extends classical operators like negation, conjunction, disjunction, and implication to handle the third value
  • Used in computer science for handling null or missing values in databases (SQL, Codd's 3VL)
• Infinite-valued logics, such as fuzzy logic, allow truth values to be any real number between 0 and 1
  • Enables representing degrees of truth and partial membership in sets
  • Applicable in control systems, decision making, and artificial intelligence (temperature control, expert systems)

Fuzzy Logic

Key concepts in fuzzy logic

• Fuzzy sets are sets with degrees of membership, allowing elements to partially belong
  • Membership functions map elements to their degree of membership, a value between 0 and 1
  • Example: a person can be "somewhat tall" with a membership of 0.7 in the fuzzy set of tall people
• Linguistic variables take on values described by natural language terms
  • Example: temperature can be "cold", "warm", or "hot" rather than just numerical values
  • Each linguistic term is associated with a fuzzy set and corresponding membership function (trapezoidal, Gaussian)
• Fuzzy operators generalize classical logic operators to work with membership functions
  • Complement (not) $\mu_{\text{not } A}(x) = 1 - \mu_A(x)$, intersection (and) $\mu_{A \cap B}(x) = \min(\mu_A(x), \mu_B(x))$, union (or) $\mu_{A \cup B}(x) = \max(\mu_A(x), \mu_B(x))$
  • Other t-norms and t-conorms can also be used to define these operators (product, Łukasiewicz, drastic product)
• Fuzzy inference systems make decisions based on fuzzy IF-THEN rules
  • Rules use linguistic variables, e.g., "IF temperature is high AND humidity is high THEN comfort is low"
  • Individual rule outputs are aggregated, defuzzified to produce a crisp output value (centroid, mean of max)

Many-valued vs classical logic

• Truth values differ between logics
  1. Classical logic uses only two truth values: true and false
  2. Many-valued logics introduce three or more discrete truth values (true, false, unknown)
  3. Fuzzy logic allows for infinitely many truth values, any real number between 0 and 1
• Ability to handle vagueness and uncertainty increases from classical to many-valued to fuzzy logic
  • Classical logic has limited capability, struggles with borderline cases and degrees of truth
  • Many-valued logics are better suited but still use discrete values rather than a continuum
  • Fuzzy logic is most adept, enabling degrees of truth and partial membership to represent vagueness
• Applications vary based on the logic system
  • Classical logic is used in mathematics, basic computer logic, and philosophical reasoning (propositional logic, Boolean algebra)
  • Many-valued logics are used in computer science, databases, and some decision-making tasks (SQL null values, circuit design)
  • Fuzzy logic is used extensively in control systems, artificial intelligence, and complex decision-making (anti-lock brakes, washing machines, risk assessment)