Skip to content

Latest commit

 

History

History
99 lines (68 loc) · 4.86 KB

README.md

File metadata and controls

99 lines (68 loc) · 4.86 KB
Raw

slsqp

Modern Fortran Edition of the SLSQP Optimizer

Status

Language GitHub release Build Status codecov last-commit

Description

This is an updated version of the SLSQP nonlinear constrained optimization code. It can be used to solve nonlinear programming problems that seek to minimize a scalar performance index subject to nonlinear equality and inequality constraints as well as bounds on the variables.

Updates to the original code include:

  • It has been translated into free-form source.
  • It is now thread safe. The original version was not thread safe due to the use of saved variables in one of the subroutines.
  • It no longer uses obsolescent and non-standard Fortran features. It should now be 100% standard compliant (Fortran 2008).
  • It now has an easy-to-use object-oriented interface. The slsqp_class is used for all interactions with the solver. Methods include initialize(), optimize(), and destroy().
  • It includes updated versions of some of the third-party routines used in the original code (BLAS, LINPACK, and NNLS).
  • Some new features were added to support printing error messages and reporting iterations to the user.
  • The user can now specify the max and min alpha to use during the line search.
  • The user can supply a routine to compute the gradients of the objective function and constraints, or allow the code to estimate them using finite differences (backward, forward, or central).
  • The documentation strings in the code have been converted to FORD format, allowing for nicely formatted documentation to be auto-generated.
  • A couple of bug fixes noted elsewhere have been applied.

Building SLSQP

Fortran Package Manager

The library can be built with the Fortran Package Manager using the provided fpm.toml file like so:

fpm build --release

By default, the library is built with double precision (real64) real values. Explicitly specifying the real kind can be done using the following processor flags:

Preprocessor flag Kind Number of bytes
REAL32 real(kind=real32) 4
REAL64 real(kind=real64) 8
REAL128 real(kind=real128) 16

For example, to build a single precision version of the library, use:

fpm build --profile release --flag "-DREAL32"

To use SLSQP within your fpm project, add the following to your fpm.toml file:

[dependencies]
slsqp = { git="https://github.com/jacobwilliams/slsqp.git" }

or, to use a specific version:

[dependencies]
slsqp = { git="https://github.com/jacobwilliams/slsqp.git", tag = "1.3.0" }

Dependencies

The library requires some BLAS routines, which are included. However, the user may also choose to link to an external BLAS library. This can be done by using the HAS_BLAS compiler directive. For example:

fpm build --compiler gfortran --flag "-DHAS_BLAS -lblas"

However, note that an external BLAS can only be used if the library is compiled with double precision (real64) reals.

Documentation

The latest API documentation can be found here. This was generated from the source code using FORD.

License

Development

  • Development continues on GitHub.

References