From Manual Rules to Automatic Prototypes: SIM as a Complete Replacement for Fuzzy Control

We show that the Similarity-Induced Method (SIM) is a complete replacement for fuzzy logic in control systems. SIM preserves the mathematical structure of fuzzy inference while eliminating all manual design: membership functions, rule tables, and parameter tuning. We identify three structural fragilities in fuzzy controllers — rule consequent errors (33–67%), membership width sensitivity (R² swings 0.48–0.99), and the resolution catastrophe (R² drops 0.93→0.25) — and show SIM is immune to all three. Validated on five benchmarks including the Tennessee Eastman industrial process. Builds on prior work DOI:10.5281/zenodo.20512632.