使用 Homebrew, apt, MacPorts, Nix 安装 lp_solve

项目链接

• 主页
• 仓库
• 软件包管理器页面

命令和别名

• lp_solve

项目历史

The official reference guide says lp_solve was originally developed by Michel Berkelaar at Eindhoven University of Technology. Jeroen Dirks' work enabled the transition from the basic 1.5 version to the fuller 2.0 version by adding a procedural interface, built-in MPS reader, fixes, and enhancements.

The same introduction records that lp_solve moved to LGPL licensing starting with version 3.0, after earlier non-commercial terms. It also says development stagnated for a time at version 3.2 before Kjell Eikland and Peter Notebaert revived development, with later version 4 and version 5 work focused on faster solving, more stability, larger models, and new functionality.

The current GitHub repository and README continue the 5.5 line. The README explains that the jump from 5.1 to 5.5 marked more than a routine update: the solver engine was revised and optimized, numerical stability improved, LUSOL became the default basis factorization package, and the code became faster and more robust than 5.1.

采用历史

lp_solve is both a standalone executable and an embeddable library, which explains its broad packaging footprint. The official introduction lists use from C, VB, .NET, Delphi, Excel, Java, AMPL, MATLAB, O-Matrix, Scilab, Octave, and R, and the batch input records Homebrew, Debian/Ubuntu development packages, MacPorts, and Nix packaging.

The official FAQ provides citation data for lp_solve as an open source mixed-integer linear programming system, with version 5.1.0.0 dated 1 May 2004 and co-developers Michel Berkelaar, Kjell Eikland, and Peter Notebaert. That reflects adoption not just as a command but as software embedded or cited in optimization work.

使用方式

The `lp_solve` command reads MIP models from standard input or files and writes results to the console or a file. The official command reference documents built-in LP and MPS input formats, external language interfaces through XLI, redirection and pipelines, and many solver options for minimization/maximization, pivoting, scaling, presolve, basis handling, branch-and-bound, tolerances, timeouts, and model conversion.

As a library, lp_solve exposes an API for building models in memory, solving them, and reading results from many programming languages. The official introduction also notes an IDE, model input through files, and model conversion facilities, while the README highlights BFP and XLI as extension points.

为什么软件包爱好者会关心

lp_solve is a classic scientific package: old, portable C code, command-line executable, shared library, examples, manuals, and language bindings. It is significant in package history because it bridges Unix solver workflows, Windows IDE use, scripting-language bindings, and academic optimization examples under an LGPL license.

时间线

• Pre-3.0: lp_solve developed by Michel Berkelaar, with version 1.5 preceding the full 2.0 transition.
• 3.0: Project starts being released under the LGPL according to the official introduction.
• 2004: Official FAQ citation data names version 5.1.0.0 dated 1 May 2004.
• 2005: README change history records beta release of version 5.5.0.0.
• 2025: `lp-solve/lp_solve` repository created on GitHub for the maintained source mirror.

Related projects

• LUSOL is the basis factorization package that became the default in the 5.5 line.
• XLI drivers connect lp_solve to external model languages such as MathProg.
• R, MATLAB, Octave, Scilab, Excel, Java, and .NET interfaces are part of the broader lp_solve ecosystem named by the official docs.

来源

• https://api.github.com/repos/lp-solve/lp_solve
• https://github.com/lp-solve/lp_solve/blob/main/README.txt
• https://lpsolve.sourceforge.net/5.5/FAQ.htm
• https://lpsolve.sourceforge.net/5.5/Intro.htm
• https://lpsolve.sourceforge.net/5.5/lp_solve.htm
• source_facts.package-manager-url