Pyomo.DoE: Add symbolic differentiation gradient calculation option as an alternative to finite different - #7

Open

adowling2 wants to merge 48 commits into

pyomo-doe-symbolic

Commits on Jun 4, 2025

Replaced FiniteDifferenceStep with GradientMethod and extended
adowling2
committed
3f32b89
Updated tests. Decided to make _gradient_method provide to discourage users from modifying it
adowling2
committed
c35c2e3
Updated reactor example to new interface. Improve solution printing. Created new temporary example for testing.
adowling2
committed
7528d7f
Added comments to plan out changes for symbolic gradients.
adowling2
committed
b97dbfa
Updated notes based on conversation with Dan
adowling2
committed
6df0470

Commits on Jun 5, 2025

Started to add logic for symbolic gradient
adowling2
committed
dcce045
Symbolic example "runs", but there are no gradient calculations yet
adowling2
committed
3ac0a52
Decided to make the kaug function also support AD
adowling2
committed
21ab51e

Commits on Jun 6, 2025

Added class for differentiation with PyNumero. Debugging matrix dimensions.
adowling2
committed
cf24135
Minor changes to examples to run on a machine without HSL installed
adowling2
committed
f791f2b
Fixed dimensions issue
adowling2
committed
fdefe5a
Added some debugging lines.
adowling2
committed
dae1c46
More debugging code
adowling2
committed
54f6769
Added some comments about next steps
adowling2
committed
49e2ddf

Commits on Jun 7, 2025

Added comments for next step
adowling2
committed
362fea9

Commits on Jun 11, 2025

Check in while developing... I think I do not actually need the output_fixed.
adowling2
committed
222430b
Intermediate check-in before removing debugging lines
adowling2
committed
aef38dd
Symbolic sensitivity analysis runs!
adowling2
committed
0eb3b9b
Generating results for comparison.
adowling2
committed
aed858b
Generating results with kaug
adowling2
committed
d45c479
Debugging with kaug
adowling2
committed
8eea498
Added simply polynomial example
adowling2
committed
0d3978b
Added scaling for symbolic
adowling2
committed
9ea42fc
Reran the heat map sensitivity analysis
adowling2
committed
9f8df60
Left some comments on next steps; enabled logging for simple timing
adowling2
committed
b8ec21c
Work in progress for symbolic version
adowling2
committed
410e6fc

Commits on Jun 12, 2025

Minor formatting change
adowling2
committed
e3bfcd5
Created suffix to track mapping
adowling2
committed
ae53caf
Updated to use measurement_mapping
adowling2
committed
92c3e47
It appears to work!
adowling2
committed
7dc978b
Need debugging help with ComponentUID
adowling2
committed
1f0e61f

Commits on Jun 13, 2025

I think it works now
adowling2
committed
cc8d579
Added scaling
adowling2
committed
1456f14

Commits on Jun 25, 2025

Began process of simplifying interface for specifying gradient methods
adowling2
committed
63852b0
Removed calls to GradientMethod.symbolic
adowling2
committed
528e31f
Added check to ensure we do not call run_doe using kaug gradient method
adowling2
committed
8f32435
Minor update to doc string
adowling2
committed
cb8826f
Added comments on where to update tests
adowling2
committed
67b7ecf
Updated examples
adowling2
committed
0574605
Reactor example now supports all five gradient modes, no need for a separate symbolic file
adowling2
committed
8545e63
Updated example to avoid depreciation warning
adowling2
committed
0ea9514
Added breadcrumbs for parameterizing tests
adowling2
committed
6b67fed
Began parameterizing tests
adowling2
committed
63cfb7c
DoE tests now all pass... still need to check code coverage
adowling2
committed
15fa4e4
Merge branch 'master' into pyomo-doe-symbolic
adowling2
committed
c62a3a8
Ran black
adowling2
committed
2fc9c8a
Fixed typos
adowling2
committed
4459a41
Removed reference to depreciated method
adowling2
committed
ce04ffa