Control Laws in Aging and Longevity

TL;DR

This paper introduces a control-theoretic framework for aging, modeling it as a loss of safe controllability, and demonstrates its utility through a five-dimensional ODE model, case studies, and empirical scoring of interventions.

Contribution

It develops a novel control-based approach to quantify and predict aging interventions, integrating safety, sequence, and modality considerations.

Findings

Control-value reduction predicts intervention success better than Hallmark annotation.

The framework generates twenty falsifiable predictions.

Empirical scoring of interventions across biological epochs supports the model.

Abstract

Aging research has produced powerful explanatory frameworks -- evolutionary theories, the Hallmarks of Aging, SENS, geroscience, hyperfunction, and information-loss models -- yet none provides quantitative rules for determining which intervention, in which biological state, at what dose, time, and sequence, will safely restore function. Existing frameworks characterize what changes during aging; they do not define equations of motion, safety constraints, or optimality conditions for drug discovery. We propose a control-theoretic framework in which aging is defined as progressive loss of safe controllability: the increasing cost and decreasing feasibility of returning a biological system to a functional viability set. Biological age is the minimum safe control cost required to restore or maintain function. Drugs are vector fields on biological state space; targets are ranked by expected cost reduction; combinations by their expansion of the reachable safe set; sequences matter because intervention vector fields do not commute. We provide a five-dimensional ODE model with analytic Lie-bracket derivation of order dependence; a modality-aware control layer distinguishing small molecules, biologics, gene therapy, and reprogramming by reversibility and safety envelope; three translational case studies (thymus, sarcopenia, ovarian aging); an implementation architecture with power analysis; and empirical scoring of 110 aging interventions across five biological epochs. The framework generates twenty falsifiable predictions. The central empirical claim is that control-value reduction predicts translational success better than Hallmark annotation or static biomarker reversal.