sparseelad / abip Goto Github PK

ABIP is a new framework that applies alternating direction method of multipliers (ADMM) to implement interior point method (IPM) for solving large-scale linear programs (LP) and conic programs (QCP).

ABIP(LP) was initially developed by Tianyi Lin (https://github.com/tyDLin) and is currently maintained by LEAVES optimization software platform.

Current version of ABIP is 2.0

v2.0 (ABIP+)

1. Added support for QCP
2. Improved LP solver

v1.0

1. ABIP for large scale linear programming problems

• Lin, Tianyi, et al. "An ADMM-based interior-point method for large-scale linear programming." Optimization Methods and Software (2020): 1-36.
• Deng, Qi, et al. "New Developments of ADMM-based Interior Point Methods for Linear Programming and Conic Programming." arXiv preprint arXiv:2209.01793 (2022).

To install ABIP, use install.m script in the main direction.

For QCP solver, user needs to specify MKL path in

mkl_path = ''; % '/opt/intel/oneapi/mkl'

ABIP accepts standard sedumi format defined using $A, b, c, K$.

To call it, use

[x, y, s, info] = abip(data, K, params)

ABIP-LP

Chenyu Xue (Phd. student at RIIS, Shanghai University of Finance and Economics. Email: [email protected])

ABIP-QCP

Jinsong Liu (Phd. student at RIIS, Shanghai University of Finance and Economics. Email: [email protected])

Tianhao Liu (Phd. student at RIIS, Shanghai University of Finance and Economics. Email: [email protected])

Testing

Chuwen Zhang (Phd. student at RIIS, Shanghai University of Finance and Economics. Email: [email protected])

Interface

Wenzhi Gao (Master student at SIME, Shanghai University of Finance and Economics. Email: [email protected])