ENTRO-CORE: A Closed-Loop Entropy-Based Control Architecture for Self-Regulated Intelligence Systems

ENTRO-CORE (E-LAB-03) is the third project of the EntropyLab research program. It introduces a closed-loop entropy control architecture that embeds thermodynamic self-regulation directly into the core operational loop of adaptive intelligence systems, advancing from the external monitoring approach of ENTRO-AI (E-LAB-02) toward intrinsic stability. The framework defines a nonlinear control law combining sigmoid and hyperbolic tangent functions acting on the entropy state Ψ(t) and its temporal derivatives (dΨ/dt, d²Ψ/dt²), with four actuation mapping strategies (Linear, Exponential, Quadratic, Threshold). Simulation validation using a synthetic damped oscillatory model demonstrates spiral attractor convergence in phase space across all actuation strategies. Builds on ENTROPIA (E-LAB-01, DOI: 10.5281/zenodo.19416737) and ENTRO-AI (E-LAB-02, DOI: 10.5281/zenodo.19284086).