DescentCore: A Verified Kernel for Descent Systems in Lean 4 — Termination and the Attractor as Theorems

Many computational and dynamical systems are organized by a single pattern: a quantity attached to each state strictly decreases at every step until it reaches zero, where the system rests. Termination of such systems is usually argued instance by instance, or assumed. This paper presents DescentCore, a small verified kernel in Lean 4 that states the pattern once, as a structure, and proves its two consequences once, as theorems: from any state, the orbit reaches an energy-zero state within a number of steps bounded by the initial energy (terminates_within), and the state reached is a rest state that the orbit never leaves (reaches_rest). No convergence axiom is assumed; convergence is derived from strict energy descent. The kernel is developed in core Lean only, without Mathlib and without sorry; its checked axiom basis is [propext, Quot.sound]. Three larger verified constructions in the same repository — a universal register machine, its stored-program form, and a stacking theorem for layers of descent systems — stand on this kernel, which supports reading the kernel as a base layer: a system built to this specification carries its termination and attractor guarantees in advance.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-14
DOI
https://doi.org/10.5281/zenodo.24215436
Citations
1
Primary Topic
Logic, programming, and type systems
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

DescentCore: A Verified Kernel for Descent Systems in Lean 4 — Termination and the Attractor as Theorems

Takeo Yamamoto
1 citations
Zenodo (CERN European Organization for Nuclear Research)
Logic, programming, and type systems
preprint

DescentCore: A Verified Kernel for Descent Systems in Lean 4 — Termination and the Attractor as Theorems

Takeo Yamamoto
preprint en
1 citations

Abstract

Many computational and dynamical systems are organized by a single pattern: a quantity attached to each state strictly decreases at every step until it reaches zero, where the system rests. Termination of such systems is usually argued instance by instance, or assumed. This paper presents DescentCore, a small verified kernel in Lean 4 that states the pattern once, as a structure, and proves its two consequences once, as theorems: from any state, the orbit reaches an energy-zero state within a number of steps bounded by the initial energy (terminates_within), and the state reached is a rest state that the orbit never leaves (reaches_rest). No convergence axiom is assumed; convergence is derived from strict energy descent. The kernel is developed in core Lean only, without Mathlib and without sorry; its checked axiom basis is [propext, Quot.sound]. Three larger verified constructions in the same repository — a universal register machine, its stored-program form, and a stacking theorem for layers of descent systems — stand on this kernel, which supports reading the kernel as a base layer: a system built to this specification carries its termination and attractor guarantees in advance.

Zenodo (CERN European Organization for Nuclear Research)
Logic, programming, and type systems
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon