verified-optimization / cvxlean Goto Github PK
View Code? Open in Web Editor NEWConvex optimization modeling in Lean 4
License: Apache License 2.0
Convex optimization modeling in Lean 4
License: Apache License 2.0
Currently, when dcp
is applied, the new conic constraints are anonymous. Instead, we could give them a more informative name. For example, new constraints coming from constraint c1
could be called c1.1
, c1.2
, etc.
Add support for the open-source SOCP solver ECOS: https://github.com/embotech/ecos.
The idea is that it takes an expression and checks its curvature using the DCP rules. Currently, one needs to define a whole problem and apply the dcp
tactic.
Currently, every rewrite step (except objective function pre-compositions) results in an application of a (custom) congruence rule. This is unnecessary. All the rewrites on a given component can be applied together.
Before: (rewrite_constraint_2
; rw_x
), (rewrite_constraint_2
; rw_y
), (rewrite_constraint_2
; rw_z
), ...
After: rewrite_constraint_2
; (rw_x
; rw_y
; rw_z
), ...
A convenient way to list all the atoms. It could also take some options, for example, to print only convex atoms.
The source of this error is still unknown. Check that we guard all calls to Expr.mvarId!
. Note that this doesn't make CvxLean crash and only appears as an error message in the process output.
Extension of pre_dcp
(for instance, something like pre_dcp only ...
) that targets only one component instead of the whole optimization problem. It is often the case that one or two constraints are problematic, but the rest are fine. So, it would be helpful for users to be able to target those.
The idea would be to do something like the Duper tutorial based on lean4web. The main challenge is that our installation has a few prerequisites, mainly Rust/Cargo to enable pre_dcp
, and MOSEK.
This can only be done properly after we solve #35 (ECOS) since MOSEK requires a license.
Initially, this can be just a version of the README file.
Ideally, do #37 first so that it can be linked.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.