Resolves #587 - Implement the Shubert-Piyavskii method for minimization #588
Conversation
Luis-Varona
force-pushed
the
587-shubertpiyavskii
branch
from
d097263 to
90d8204
Compare
|
Just added a minor bugfix to one of the doctests (didn't affect the actual implementation, though)--squashed into the same commit to avoid clutter. |
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #588 +/- ##
==========================================
+ Coverage 92.13% 92.17% +0.04%
==========================================
Files 177 178 +1
Lines 23672 24139 +467
==========================================
+ Hits 21810 22250 +440
- Misses 1862 1889 +27 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
Luis-Varona
force-pushed
the
587-shubertpiyavskii
branch
from
90d8204 to
c99ead9
Compare
|
One last bugfix--a spelling correction ("arbitraily" -> "arbitrarily") in the examples file and making the PartialOrd impl for SearchInterval more idiomatic in the main solver implementation. Should pass all CI tests now! 🙂 |
Luis-Varona
force-pushed
the
587-shubertpiyavskii
branch
from
c99ead9 to
4bf9bf2
Compare
The Shubert-Piyavskii method optimizes a univariate, Lipschitz continuous function inside a specified interval. Given an accurate Lipschitz constant, it is guaranteed to come within an arbitrary epsilon of the true global minimum. Unlike most other deterministic global optimizations algorithm (e.g., golden-section search), it is not restricted to unimodal functions; moreover, it does not require an initial guess. It deterministically samples points from ever-narrowing subintervals of the search space until the sampled best point is sufficiently close to the lower bound given by the Lipschitz constant. I have added an implementation of an SP solver in crates/argmin/src/solver/shubertpiyavskii/, modifying crates/argmin/src/lib.rs and crates/argmin/src/solver/mod.rs accordingly. Additionally, I have added an example run of the solver to examples/shubertpiyavskii/.
Luis-Varona
force-pushed
the
587-shubertpiyavskii
branch
from
4bf9bf2 to
5d75b8d
Compare
|
One last fix; code and doctests were all working, but clippy got mad at something non-idiomatic in a match branch, hehe. Now it should be 100% ready, pending your approval @stefan-k :) (No rush!) |
|
Hi @stefan-k! Definitely no rush, but I just wanted to ask if you still planned to get around to reviewing this at some point in the future (doesn't matter when!) 🙂 I also plan to make a PR sometime soon for Differential Evolution as per my conversation with @steinhauserc in #499, if you think that's a good idea. |
The Shubert-Piyavskii method optimizes a univariate, Lipschitz continuous function inside a specified interval. Given an accurate Lipschitz constant, it is guaranteed to come within an arbitrary epsilon of the true global minimum. Unlike most other deterministic global optimizations algorithm (e.g., golden-section search), it is not restricted to unimodal functions; moreover, it does not require an initial guess. It deterministically samples points from ever-narrowing subintervals of the search space until the sampled best point is sufficiently close to the lower bound given by the Lipschitz constant.
I have added an implementation of an SP solver in crates/argmin/src/solver/shubertpiyavskii/, modifying crates/argmin/src/lib.rs and crates/argmin/src/solver/mod.rs accordingly. Additionally, I have added an example run of the solver to examples/shubertpiyavskii/.