Releases: mdolab/pyoptsparse
Releases · mdolab/pyoptsparse
v2.4.1
pyOptSparse v2.4.1 release notes
This is a patch release that fixes several bugs:
- changed SLSQP output file id (#198)
- fixed SNOPT workspace lengths (#197)
In particular, #197 fixes a critical bug introduced in v2.4.0 that prevents certain large optimizations from running with SNOPT.
v2.4.0
pyOptSparse v2.4.0 release notes
This is a minor release that adds mdolab-baseclasses
as a dependency. It should be automatically pip-installed from PyPI if it is not available locally.
No other user-facing changes are expected, but see #193 for full details. To be clear, external optimizers (SNOPT, IPOPT and ParOpt) do not have hard-coded default values (and have never had them hard-coded), but this is now explicitly stated. We do not maintain an options list so please refer to the documentation for each solver for details. For these optimizers, there is no options checking so it is up to each optimizer to decide how to handle unrecognized or invalid option names/values. This is again the same behaviour as before, but we have removed the misleading options table on the documentation site since we do not use those information.
Other minor changes:
- Documentation updates (#186, #188)
- Added missing offset attribute to integer dvs (#192)
v2.3.0
pyOptSparse v2.3.0 release notes
This release adds support for SNOPT v7.7.5 (#185), including two new features:
Backoff factor
Proximal iterations limit
v2.2.2
pyOptSparse v2.2.2 release notes
Highlights:
- Non-objective/constraint values stored in the history file can now be accessed via
getValues
(#184)
- pyOptSparse now has a journal paper! The citation information has been updated (#178)
- Various maintenance updates
v2.2.1
pyOptSparse v2.2.1 release notes
This is a patch release primarily focused on documentation updates.
v2.2.0
pyOptSparse v2.2.0 release notes
Backwards-incompatible changes
- Python 2 is now no longer supported (#164). We currently run tests using Python 3.7 and 3.8, but other versions of Python 3 may also work.
- The flat, OpenMDAO-style sensitivity dictionary is no longer supported (#165). Please use the nested approach from now on.
v2.1.7
pyOptSparse v2.1.7 release notes
This is another patch release for pyOptSparse. Changes are:
- Fixed bug for unconstrained optimization with PSQP, CONMIN, and NLPQLP (#162)
- Documentation update (#161)
- Removed patch version from dependencies (#160)
v2.1.6
pyOptSparse v2.1.6 release notes
This is another patch release for pyOptSparse. Highlights:
- Fixed installation bug for IPOPT (#153)
- Updated documentation and other maintenance tasks
v2.1.5
pyOptSparse v2.1.5 release notes
This releases fixes the following bug:
- Added negative informs to PSQP (#148)
v2.1.4
pyOptSparse v2.1.4 release notes
This is another bugfix release for pyOptSparse.
Bug fixes
- Various fixes to the History object (#140, #142, #144)
Maintenance:
- Updated
setup.py
to make optview
an entry-point available as a CLI program without needing an alias (#134)