1 Multifrontal sparse symmetric linear solver
- Designed for: sparse, symmetric, indefinite systems.
- Efficient for sparse KKT-systems from constrained optimization, saddle point systems and similar from DAEs Differential Algebraic Equations and similar.
- Single-core implementation. Thread-based paralellization planned.
- Download: liblindefsol.so lindefsol.h
- Tested on Ubuntu 20.04, uses minimal system calls, should work on almost any modern Linux distribution.
2 Nonlinear constrained optimization solver compiler
Based on a nonlinear interior-point penalty method with the multifrontal sparse symmetric linear solver as its main workhorse.
Key characteristics: * Optimized for fast solutions to medium sized problems, eg. real-time optimal control.
- Create nonlinear model description using a
- Compiler takes as input the
.jsonfile, and outputs an optimized nonlinear solver for the given problem.
The optimized linear solver is output as a C-library, with a standard C-interface.
- Download: liboptimizer.so optimizer.h
Tested on Ubuntu 20.04, uses minimal system calls, should work on almost any modern Linux distribution.