Class Logic

constructor

• new Logic(): Logic

  Returns Logic

c

init

• init(output: string, shape: Fuzzifier): this

  Registers the first rule. Must be called before and, or, not or defuzzify.

  Rule composition is sequential and order-dependent: at defuzzify time each rule's membership is folded into a single value that is carried to the next rule (see and). init and or start a fresh value from the rule's own shape; and narrows it; not inverts it.

  Parameters

  • output: string

    Output label, letters only (/^[a-z]+$/i), must be unique.

  • shape: Fuzzifier

    Fuzzifier mapping a crisp input to a membership degree 0..1.

  Returns this

and

• and(output: string, shape: Fuzzifier): this

  Narrows the carried membership with min(previous, shape(value)) (the standard fuzzy-AND / T-norm). Because previous is whatever the immediately preceding rule produced, and is order-dependent — it composes against the prior rule in the chain, not a grouped rule set.

  Parameters

  • output: string

    Output label, letters only, must be unique.

  • shape: Fuzzifier

    Fuzzifier mapping a crisp input to a membership degree 0..1.

  Returns this

or

• or(output: string, shape: Fuzzifier): this

  Starts a fresh membership from this rule's own shape value, ignoring the value carried from previous rules (fuzzy-OR / max is applied implicitly by defuzzify picking the single highest membership across all rules).

  Parameters

  • output: string

    Output label, letters only, must be unique.

  • shape: Fuzzifier

    Fuzzifier mapping a crisp input to a membership degree 0..1.

  Returns this

not

• not(output: string, shape: Fuzzifier): this

  Inverts this rule's shape value: 1 - shape(value) (fuzzy-NOT / complement). Like or, it does not depend on the carried value.

  Parameters

  • output: string

    Output label, letters only, must be unique.

  • shape: Fuzzifier

    Fuzzifier mapping a crisp input to a membership degree 0..1.

  Returns this

defuzzify

• defuzzify(value: number, as?: string): DefuzzifyResult

  Evaluates every rule for value and returns the result.

  Note on terminology: this performs max-membership classification, not centroid / center-of-gravity defuzzification. It returns the label of the rule with the highest membership (defuzzified) together with that membership degree (fuzzified) — it does not compute a crisp output number. On ties the earliest rule wins (strict greater-than).

  Parameters

  • value: number

    Crisp input fed to every rule's shape.

  • Optionalas: string

    Optional namespace prefix for DefuzzifyResult.boonJsInputs keys, e.g. 'heat' yields heat.cold. Useful when merging the inputs of several Logic instances for boon-js evaluation.

  Returns DefuzzifyResult

  Example

  logic.defuzzify(10)
  Copy