Primal-dual interior point method for hierarchical least-squares programs with linear constraints

This software provides an interior-point method for hierarchical least-squares programs with linear constraints. The corresponding paper can be found here "NIPM-HLSP: An Efficient Interior-Point Method for Hierarchical Least-Squares Programs" ( https://arxiv.org/pdf/2106.13602.pdf ).

NIPM-HLSP solves hierarchical least-squares problems of the form

min. x: ||A_l x - b_l||_2^2 , l = 1,...,p

s.t \underline{A}{l-1} x - \underline{b}{l-1} = w_{l-1}^*

A_l are dense matrices of any form and can be rank-deficient.

Install required packages

Please make sure to install the following packages:

• Eigen >= 3.2.10

Getting started

Clone the repository to the desired work folder <work_folder>

cd <work_folder>
mkdir build
cd build
cmake ..
ccmake .

set CMakeBuildType to Release set COMPILE_TESTS to OFF

make && sudo make install
ccmake .

set COMPILE_TESTS to ON

make && sudo make install

run ./build/tests/test1

License

Copyright (c) 2022, Kai Pfeiffer

Authors

• Kai Pfeiffer (kaipfeifferrobotics@gmail.com)