Back to Search Start Over

Verified reductions for optimization

Authors :
Bentkamp, Alexander
Mir, Ramon Fernández
Avigad, Jeremy
Publication Year :
2023

Abstract

Numerical and symbolic methods for optimization are used extensively in engineering, industry, and finance. Various methods are used to reduce problems of interest to ones that are amenable to solution by such software. We develop a framework for designing and applying such reductions, using the Lean programming language and interactive proof assistant. Formal verification makes the process more reliable, and the availability of an interactive framework and ambient mathematical library provides a robust environment for constructing the reductions and reasoning about them.

Details

Database :
arXiv
Publication Type :
Report
Accession number :
edsarx.2301.09347
Document Type :
Working Paper