Separate Linear Terms in Nonlinear to Piecewise Linear Transformation by michaelbynum · Pull Request #3814 · Pyomo/pyomo

michaelbynum · 2026-01-13T15:11:54Z

Summary/Motivation:

This PR modifies the contrib.piecewise.nonlinear_to_pwl transformation so that linear parts of a constraint always get separated from the nonlinear parts, even if additively_decompose is False. The idea is that

y >= exp(x) + x**3

should become

y >= PWL(exp(x) + x**3)

and not

0 >= PWL(exp(x) + x**3 - y)

Legal Acknowledgement

By contributing to this software project, I have read the contribution guide and agree to the following terms and conditions for my contribution:

I agree my contributions are submitted under the BSD license.
I represent I am authorized to make the contributions and grant the license. If my employer has rights to intellectual property that includes these contributions, I represent that I have received permission to make contributions and grant the required license on behalf of that employer.

codecov · 2026-01-13T16:15:28Z

Codecov Report

❌ Patch coverage is 97.77778% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 89.93%. Comparing base (52a29ae) to head (8431367).
⚠️ Report is 17 commits behind head on main.

Files with missing lines	Patch %	Lines
...mo/contrib/piecewise/transform/nonlinear_to_pwl.py	97.77%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3814      +/-   ##
==========================================
- Coverage   89.94%   89.93%   -0.01%     
==========================================
  Files         902      902              
  Lines      106354   106393      +39     
==========================================
+ Hits        95655    95680      +25     
- Misses      10699    10713      +14

Flag	Coverage Δ
builders	`29.19% <11.11%> (-0.01%)`	⬇️
default	`86.23% <97.77%> (?)`
expensive	`35.64% <11.11%> (?)`
linux	`87.38% <97.77%> (-2.04%)`	⬇️
linux_other	`87.38% <97.77%> (-0.01%)`	⬇️
oldsolvers	`28.10% <11.11%> (-0.01%)`	⬇️
osx	`82.71% <97.77%> (-0.01%)`	⬇️
win	`85.80% <97.77%> (-0.02%)`	⬇️
win_other	`85.80% <97.77%> (-0.02%)`	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

emma58

I think this makes sense. If you could put in a comment with the example you put in the PR description for why _separate_linear_parts is necessary, I think that would be helpful for future us. My guess is this is not currently the fastest thing we could possibly do, but for now, I'm not so concerned with that. @jsiirola, @sadavis1, @bammari, this might be worth glancing at if you have a minute.

emma58 · 2026-01-13T16:29:00Z

pyomo/contrib/piecewise/transform/nonlinear_to_pwl.py

+                    if x1 == x2:
+                        nonlinear += coef * var_map[x1] ** 2
+                    else:
+                        nonlinear += coef * (var_map[x1] * var_map[x2])


Does separating these two cases matter? I mean, obviously you make a different expression tree, but do we need it?

I don't think it matters to me.

I'm going to leave this just for consistency with QuadraticRepn.

emma58 · 2026-01-13T16:30:14Z

pyomo/contrib/piecewise/transform/nonlinear_to_pwl.py

+        if repn.multiplier_flag(repn.multiplier) != 1:
+            linear *= repn.multiplier
+            nonlinear *= repn.multiplier


@jsiirola, can this happen? I have something in the back of my mind telling me multiplier is sure to be 1 at this point?

I just copied the to_expression code from the QuadraticRepn class and modified it slightly...

which is to say that I have no idea. I'll take a look, though.

Oh, I can't read. This question was for @jsiirola. My bad.

michaelbynum · 2026-01-13T18:41:40Z

Okay, if this does make some sense, I'll clean it up a bit.

michaelbynum · 2026-01-20T16:00:09Z

Okay, assuming tests pass, this should be good to go.

emma58

This looks good to me. I still have the multiplier question, if you can check that when you look, @jsiirola

pyomo/contrib/piecewise/transform/nonlinear_to_pwl.py

michaelbynum · 2026-02-17T20:21:20Z

pyomo/contrib/piecewise/transform/nonlinear_to_pwl.py

+            linear *= repn.multiplier
+            nonlinear *= repn.multiplier
+
+        return linear, nonlinear


let's not lump the quadratic part back into the nonlinear part. additively decompose will separate it again.

I started to do this, but the min_dimension_to_additively_decompose option made it more complex than I thought was worthwhile.

I'm taking back what I said. I'm going to make this change. I was confused.

Okay, this is done, now.

michaelbynum · 2026-03-23T14:27:12Z

Assuming tests pass, this should be ready for review.

michaelbynum · 2026-03-24T20:13:19Z

So this is not ready for review yet.

michaelbynum · 2026-03-25T12:08:56Z

Assuming tests pass, this should be ready for review.

jsiirola

A couple minor questions / suggestions....

jsiirola · 2026-03-25T17:37:25Z

pyomo/contrib/piecewise/transform/nonlinear_to_pwl.py

+        0 >= PWL(exp(x) + x**3 - y)
+        """
+        assert repn.multiplier == 1
+        linear_repn = QuadraticRepn()


If this will always just have constant + linear, why not use LinearRepn?

It feels weird to use a LinearRepn with a QuadraticRepnVisitor?

Maybe. But I suppose no weirder than

linear_repn = QuadraticRepn()

So true... I should not be allowed to name things.

Me neither....

pyomo/contrib/piecewise/transform/nonlinear_to_pwl.py

michaelbynum added 2 commits January 13, 2026 08:04

always separate linear parts in nonlinear to pwl transformation

02daec1

always separate linear parts in nonlinear to pwl transformation

7a521ab

michaelbynum requested a review from emma58 January 13, 2026 15:11

Merge remote-tracking branch 'origin/main' into pwl_approx

7b78841

emma58 reviewed Jan 13, 2026

View reviewed changes

michaelbynum added 4 commits January 20, 2026 08:25

fix piecewise tests

b16b0b6

run black

f3b8936

add test to piecewise

e2d6264

Merge remote-tracking branch 'origin/main' into pwl_approx

dd2dbe4

emma58 approved these changes Jan 20, 2026

View reviewed changes

blnicho requested a review from jsiirola January 21, 2026 06:19

blnicho added this to Pyomo 6.10 Jan 21, 2026

github-project-automation bot moved this to Todo in Pyomo 6.10 Jan 21, 2026

github-project-automation bot moved this from Todo to Reviewer Approved in Pyomo 6.10 Jan 21, 2026

blnicho moved this from Reviewer Approved to Review In Progress in Pyomo 6.10 Jan 21, 2026

Merge branch 'main' into pwl_approx

cff61ab

blnicho removed this from Pyomo 6.10 Feb 10, 2026

pyomo-autotest added the AT: STALE label Feb 11, 2026

michaelbynum commented Feb 17, 2026

View reviewed changes

pyomo/contrib/piecewise/transform/nonlinear_to_pwl.py Outdated Show resolved Hide resolved

michaelbynum commented Feb 17, 2026

View reviewed changes

michaelbynum added 3 commits March 23, 2026 07:57

Merge remote-tracking branch 'origin/main' into pwl_approx

287e47a

simplify logic for separating linear and nonlinear parts

466e6fa

run black

7cc648a

pyomo-autotest removed the AT: STALE label Mar 23, 2026

Merge remote-tracking branch 'origin/main' into pwl_approx

b4c4cac

michaelbynum added 2 commits March 25, 2026 06:07

avoid extra work

9852beb

run black

a708a8d

edge cases

1b547ce

jsiirola reviewed Mar 25, 2026

View reviewed changes

cleanup

8431367

jsiirola approved these changes Mar 25, 2026

View reviewed changes

michaelbynum merged commit 6ab0106 into Pyomo:main Mar 26, 2026
64 of 65 checks passed

Conversation

michaelbynum commented Jan 13, 2026

Summary/Motivation:

Legal Acknowledgement

Uh oh!

codecov bot commented Jan 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

emma58 left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

emma58 Jan 13, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

michaelbynum commented Jan 13, 2026

Uh oh!

michaelbynum commented Jan 20, 2026

Uh oh!

emma58 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

michaelbynum commented Mar 23, 2026

Uh oh!

michaelbynum commented Mar 24, 2026

Uh oh!

michaelbynum commented Mar 25, 2026

Uh oh!

jsiirola left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

codecov bot commented Jan 13, 2026 •

edited

Loading

emma58 Jan 13, 2026 •

edited

Loading