A Lean 4 formalization of the Minimum Asymmetry Gap (MAG) framework, which translates P vs NP into group theory.
MAG provides a group-theoretic semantics where:
| Concept | MAG Translation |
|---|---|
| Polytime algorithm | Solvable group |
| NP-hard problem | Non-solvable group (A₅ embedding) |
| Solves | Group isomorphism |
The core insight: solvable groups cannot be isomorphic to non-solvable groups, yielding P ≠ NP within this semantics.
| Theorem | Description | Status |
|---|---|---|
| Theorem 3.1 | MAG internal P ≠ NP | ✅ sorry-free |
| Theorem 4.6 | A₅ Universal Barrier | ✅ sorry-free |
| Theorem 4.7 | Solvable vanishing depth | ✅ sorry-free |
| Theorem 5.4 | Bridge theorem (conditional) | ✅ sorry-free |
| Theorem 5.7 | Toy model verification | ✅ sorry-free |
# Build core theorems (default)
lake build
# Build support modules
lake build Support
# Build all (including blueprints)
lake build MAG- Lean 4.27.0-rc1
- Mathlib
This formalization does NOT unconditionally claim to solve the Clay P vs NP problem.
- Unconditional: MAG-internal P ≠ NP (group-theoretic definitions)
- Conditional: If
TranslationInterfaceholds, then standard P ≠ NP
MAG synthesizes:
- Galois (1832): Solvability concept
- Klein (1872): Erlangen Program methodology
- Krohn-Rhodes (1965): Automata decomposition
- Barrington (1986): NC¹ = width-5 BP over S₅
- CFSG: A₅ as minimal non-solvable simple group
MIT License
Masaru Numagaki, January 2026