Compositional polynomial-time computation with black-box subroutines
No public endorsements yet.
Loading review…
Sign in with ORCIDAbstract
Polynomial-time algorithms can be combined using implementations supplied only by existence theorems. We give a structured stack language with calls to a fixed library of input/output contracts and compile it to ordinary finite stack Turing machines. Calls preserve the caller's other stacks and finite control, and reset all private storage. Their exact cost is the callee's running time plus two traversals of each argument and result and four transfer transitions.
Polynomial bounds on source execution and on call argument and result sizes give polynomial bounds on the compiled machine. The resulting interface supports sequencing, conditionals, loops, and adaptive calls inside verified programs. As a direct public corollary, encoded polynomial-time functions are closed under sequential composition, even when their implementations are supplied only by existence theorems.
Concepts
Concept map
Proofs
Proof networkview on GitHub
Proof list
Lean sources for these proofs: proofs/ on GitHub
Proof code is not displayed; the archive records each proof's checked relationship between claims.
Related submissions
No other submission in the archive builds on this one, and this one builds on none.
Cite this
This is only the formalizers. The authors of the formalized results may be different (see References).
@misc{lax-218471,
author = {Szymon Toruńczyk and GPT},
title = {Compositional polynomial-time computation with black-box subroutines},
year = {2026},
howpublished = {Lax Archive, lax-218471},
url = {https://laxarchive.org/lax-218471/},
}
Discussion
Ask a question or add context. Endorsements and structured flags are kept in the review panel above.
0 comments