sinha-abhi / PDIPS.jl

Homogeneous self-dual interior point linear optimization solver

Primal-Dual Interior-Point Solver (PDIPS)

PDIPS (paper) is an implementation of the homogeneous self-dual interior-point algorithm for solving linear optimization problems using interior-point methods.

This implementation solves linear programs of the form

min     c'x
 x
s.t    Ax = b
    lo <= x <= hi

by transforming it into standard form:

min     c'x
 x
s.t    Ax = b
       0 <= x.

Implemented for the final project of Computational Methods in Optimization with David Gleich at Purdue.

Reference

For a detailed explanation (and a more general implementation) of the algorithm see Tulip.jl.