Skip to content

Releases: OCamlPro/ocplib-simplex

0.5.1

28 Mar 15:15
12e7c1f
Compare
Choose a tag to compare

CHANGES:

  • Add documentation for solving system (PR #16).
  • Separate types for coefficents and values (PR #17).
  • Remove the dependency on num (PR #19).
  • Remove messages at the App level (PR #22).

Version 0.4.1

21 Apr 11:48
Compare
Choose a tag to compare

version 0.4.1, April 21, 2023

  • fix the issue #13

Version 0.5

15 Nov 10:09
9b0944c
Compare
Choose a tag to compare
  • Reworking the library build system, now only relying on dune.
    The Makefile is now clearer and simpler to use.

  • (!!!) Logs are handled by the logs library and debug is activated by this
    library.

  • (!!!) The Rat2 module now abstract bounds as strict upper, strict lower or
    soft bounds instead of pairs of rationals.

Version 0.4

23 Aug 13:03
Compare
Choose a tag to compare
  • now, asserting bounds returns whether these bounds are trivially implied by those that are already known ( /!\ this may break code that uses previous versions )

  • add a field nb_pivots in the environment to count the number of pivots that have been made so far

  • some other optimizations

Release 0.3

09 Nov 12:19
Compare
Choose a tag to compare

Changes w.r.t. version 0.2:

  • bugfix in maximization

New release (0.2) that supports linear optimization

24 Aug 10:16
Compare
Choose a tag to compare

Changes w.r.t. version 0.1:

  • add support for linear optimization
  • some bugfixes when assuming inconsistent bounds
  • improve build and testing

First release of the library (v 0.1)

11 Jul 16:49
Compare
Choose a tag to compare

Fully functional implementation with:

  • the ability to handle strict inequalities over rationals
  • heavy debug mode to check a range of invariants:
    • before / after assuming bounds on variables
    • before / after assuming bounds on polynomials
    • after each solving step.
  • incremental and backtrackable capabilities
  • production of unsat-cores when unsatisfiable (and solutions when satisfiable)

Currently, linear optimization is not supported.