QP a general convex qpp solver

版本 1.1.0.0 (12.7 KB) 作者: Alain Barraud
QP solves dense convex quadratic programming and supports ill conditionned problem
5.0
(1)
205.0 次下载
更新时间 2023/1/31

查看许可证

QP is a self contained quadratic programming solver based upon the well known Goldfarb Idnani algorithm. Hessian factorization and orthogonal transformation are used every where (Householder and givens rotation) it is necessary. When equality constraints exist, only the reduced Hessian is required to be positive definite. The full Hessian matrix is checked against a possibly very small negative eigenvalue due to rounding errors.
The calling sequence is similar to quadprog except that there is no starting point x0 argument. Option is limited to a tolerance tol and a maximum of iterations. The exitflag convention is maintained for >=0 value and is little bit different for <0 values. The output.status gives explanation as usually. A field .drop shows how many times a previously added constraint has been dropped.
Fixed variables are set through lb(*)=ub(*). Equality constraints may be linear dependent and they are checked against incompatibilty.
The cpu time is similar to quadprog and precision is usually better in particular for ill conditionned problem.
The package has been written with matlab R2018a however it does run successfully with many older releases.

引用格式

Alain Barraud (2024). QP a general convex qpp solver (https://www.mathworks.com/matlabcentral/fileexchange/67864-qp-a-general-convex-qpp-solver), MATLAB Central File Exchange. 检索来源 .

MATLAB 版本兼容性
创建方式 R2018a
兼容任何版本
平台兼容性
Windows macOS Linux
类别
在 Help Center 和 MATLAB Answers 中查找有关 Quadratic Programming and Cone Programming 的更多信息

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

GoldfarbIdnani

版本 已发布 发行说明
1.1.0.0

correction of a bug when there is no equality constraint and modification of test to check non strictly convex problem
1.0.0.0