Thermodynamic-Complexity Duality: Embedding Computational Hardness as a Thermodynamic Coordinate

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Neukart, Florian, Vinokur, Valerii
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929687996923904
author Neukart, Florian
Vinokur, Valerii
author_facet Neukart, Florian
Vinokur, Valerii
contents We propose a duality between thermodynamics and computational complexity, elevating the difficulty of a computational task to the status of a thermodynamic variable. By introducing a complexity measure C as a novel coordinate, we formulate an extended first law, dU = T dS - p dV + ... + lambda dC, capturing energy costs beyond classical bit erasures. This perspective unifies ideas from Landauer's principle with the combinatorial overhead of hard (e.g., NP-complete) problems, suggesting that algorithmic intractability can manifest as an additional contribution to thermodynamic potentials. We outline how this "complexity potential" might produce phase-transition-like signatures in spin glasses, random constraint satisfaction, or advanced computing hardware near minimal dissipation. We also discuss parallels with previous geometry-information dualities, emphasize the role of complexity in shaping energy landscapes, and propose experimental avenues (in reversible computing or spin-glass setups) to detect subtle thermodynamic signatures of computational hardness. This framework opens a route for systematically incorporating complexity constraints into physical modeling, offering a novel link between the fundamental cost of computation and thermodynamic laws.
format Preprint
id arxiv_https___arxiv_org_abs_2501_15950
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Thermodynamic-Complexity Duality: Embedding Computational Hardness as a Thermodynamic Coordinate
Neukart, Florian
Vinokur, Valerii
General Physics
We propose a duality between thermodynamics and computational complexity, elevating the difficulty of a computational task to the status of a thermodynamic variable. By introducing a complexity measure C as a novel coordinate, we formulate an extended first law, dU = T dS - p dV + ... + lambda dC, capturing energy costs beyond classical bit erasures. This perspective unifies ideas from Landauer's principle with the combinatorial overhead of hard (e.g., NP-complete) problems, suggesting that algorithmic intractability can manifest as an additional contribution to thermodynamic potentials. We outline how this "complexity potential" might produce phase-transition-like signatures in spin glasses, random constraint satisfaction, or advanced computing hardware near minimal dissipation. We also discuss parallels with previous geometry-information dualities, emphasize the role of complexity in shaping energy landscapes, and propose experimental avenues (in reversible computing or spin-glass setups) to detect subtle thermodynamic signatures of computational hardness. This framework opens a route for systematically incorporating complexity constraints into physical modeling, offering a novel link between the fundamental cost of computation and thermodynamic laws.
title Thermodynamic-Complexity Duality: Embedding Computational Hardness as a Thermodynamic Coordinate
topic General Physics
url https://arxiv.org/abs/2501.15950