3.0.0

• Update and refinement of the bounds_mode option. There are now four options:
  • Ignore bounds (default)
  • Scalar
  • Vector
  • Wall
• Added a new option norm_mode to specify how to compute the norm of the function vector to check for convergence. There are three options:
  • 2-norm (default)
  • infinity-norm
  • 1-norm
• Updates to readme, docstrings, and tests.

2.3.0

• Added a new bounds_mode option. Bounds can either be ignored (original behavior) or a crude method used where x is adjusted at each function evaluation so that each individual element is within the specified bounds.
• CI, Readme, and unit test updates.

2.2.0

• The user can now provide a custom routine to solve the sparse linear system.

2.1.0

• Added a sparse Broyden option
• Added the LSMR solver as an option to solve the linear system.
• only print Minimum alpha reached message when verbose=True
• Readme updates
• Doc string updates

2.0.0

• Added new options for using sparse solvers (LSQR or LUSOL) for the linear system.
• CI and Readme updates

1.1.1

• No code changes, just CI and Readme updates

1.1.0

• Can now specify the real kind with a preprocessor flag (e.g., REAL32, REAL64, or REAL128).
• The Fmin library is now a dependency in FPM.

1.0.0

Initial release