Module fuzzy 

Structs

MamdaniRule

One Mamdani rule’s contribution: its firing strength and its consequent membership function consequent_mu sampled over the shared output universe.

SugenoRule

One Sugeno (TSK) rule: its firing strength and a crisp consequent value (a 0th-order constant, or a pre-evaluated 1st-order linear function of the inputs).

Functions

conjunction

Fuzzy conjunction (Gödel t-norm = min) of the truth degrees carried by quins. Empty input → 1.0 (the t-norm identity). Zero-heap.

defuzz_bisector

Bisector: the universe point that splits the area under μ into two equal halves.

defuzz_centroid

Centroid / centre-of-gravity (COG): Σ(uᵢ·μᵢ) / Σ μᵢ.

defuzz_mean_of_max

Mean-of-Maximum (MOM): the mean of the universe points attaining maximum membership.

defuzz_smallest_of_max

Smallest-of-Maximum (SOM): the smallest universe point attaining maximum membership.

degree

Read a proposition’s fuzzy truth degree (canonical f32 in metadata, via the FrameLayout ABI), clamped to [0,1].

disjunction

Fuzzy disjunction (Gödel t-conorm = max) of the truth degrees carried by quins. Empty input → 0.0 (the t-conorm identity). Zero-heap.

firing_strength

Rule firing strength = Gödel t-norm (min) of the antecedent membership degrees. Empty antecedent → 1.0 (the t-norm identity). Use degree() to source each membership from a Quin.

fuzzy_not

Standard fuzzy negation (complement) — 1 - a, clamped to [0,1].

hedge_extremely

Concentration hedge “extremely” — μ³.

hedge_more_or_less

Dilation hedge “more or less” / “somewhat” — √μ (broadens, raises partial memberships).

hedge_very

Concentration hedge “very” — μ² (sharpens, lowers partial memberships).

mamdani_infer

Mamdani inference: clip each rule’s consequent at its firing strength (min-implication), aggregate across rules by max into scratch, then defuzzify by centroid over universe. None if the aggregate set carries no mass. Zero-heap (caller owns scratch, sized to the universe).

sugeno_infer

Sugeno (TSK) inference: the firing-strength-weighted average of rule consequents, Σ(wᵢ·zᵢ) / Σ wᵢ. None if total firing is ~0 (refuse rather than divide by zero).

t_conorm_drastic

Drastic t-conorm — the largest t-conorm: b if a==0, a if b==0, else 1.

t_conorm_godel

Gödel t-conorm (fuzzy OR) — the maximum.

t_conorm_lukasiewicz

Łukasiewicz t-conorm (fuzzy OR) — min(1, a + b).

t_conorm_product

Product t-conorm (fuzzy OR) — a + b - a·b (the probabilistic sum).

t_norm_drastic

Drastic t-norm — the smallest t-norm: b if a==1, a if b==1, else 0.

t_norm_godel

Many-valued / fuzzy logic over truth degrees in [0, 1]. Distinct from the Bayesian probabilistic modality: fuzzy conjunction uses a t-norm (not a product), modelling DEGREES of (partial) satisfaction — e.g. a right that is partially fulfilled. Each proposition carries its truth degree as an f32 in the quin metadata. Zero-heap throughout. Gödel t-norm (fuzzy AND) — the minimum.

t_norm_lukasiewicz

Łukasiewicz t-norm (fuzzy AND) — max(0, a + b - 1).

t_norm_product

Product t-norm (fuzzy AND) — a · b (the probabilistic/algebraic conjunction).