JLPDEMO

This code demonstrates the progress of an interior-point method in solving a simple linear program in two variables. Select strictly feasible starting points for x and y and click the “Solve” button. Successive Unconstrained Minimization Technique (SUMT) is used to solve the linear program and the path taken by the solver is plotted. Note that the algorithm parameters were tweaked to make the plots a little more interesting. Please be patient while the Java files are downloading.

Your browser does not appear to support one or more features of this applet, sorry!

The solver itself can be used to solve medium sized problems (this is NOT a high performance solver). There’s an abstract base class called SUMT and a derived class SUMTlp that’s used to solve the linear program. You are welcome to download and use any portion of JLPDEMO including the solver. You will need to download and install JLapack, a version of BLAS and Lapack for Java. The on-line documentation will help you understand the code. There are also links below for downloading more information.

The SUMTlp class solves the primal linear program

and the corresponding dual linear program

given a strictly feasible starting point x⁰ for the primal linear program.

Downloads:

Ø Binaries (.class files)

Ø Documentation

Ø Source (.java files)

Please share any improvements you make by sending me an email!