[Ipopt-tickets] [Ipopt] #60: Support for the Trilinos Amesos sparse linear solver library

[Ipopt]

#60: Support for the Trilinos Amesos sparse linear solver library
------------------------------+---------------------------------------------
Reporter:  nowozin            |       Owner:  ipopt-team
    Type:  enhancement        |      Status:  new       
Priority:  normal             |   Component:  Ipopt     
 Version:  3.3 (C++ Version)  |    Severity:  normal    
Keywords:                     |  
------------------------------+---------------------------------------------
 MUMPS, the only liberally licensed sparse linear system solver currently
 supported by IpOpt is not as robust as the supported commercial
 alternatives (PARDISO, MA27).

 The Trilinos project produces high quality numerical software and has a
 linear system solver library called Amesos:
 http://trilinos.sandia.gov/packages/amesos/

 The library contains interfaces to PARDISO, SuperLU and MUMPS among others
 for sparse matrices and LAPACK for dense matrices (with transparent
 conversion from sparse to dense), but also contains its own solver KLU,
 which was originally the precursor to SuperLU.  All solvers in the library
 support the same interface which allows symbolic factorization (given only
 the matrix structure) and repeated numerical solution.  The library itself
 and KLU is licensed under the LGPL and thus would allow distribution of
 binary precompiled IpOpt packages.

 This ticket asks whether KLU/Amesos can be considered for inclusion into
 IpOpt, hopefully allowing a free distribution of a powerful linear solver
 together with IpOpt and at the same time supporting dense matrices well.

 The interface is described on an abstract level at
 http://trilinos.sandia.gov/packages/amesos/AmesosDesign.pdf
 and a guide to installing and using Amesos is at
 http://trilinos.sandia.gov/packages/amesos/AmesosReferenceGuide.pdf

-- 
Ticket URL: <https://projects.coin-or.org/Ipopt/ticket/60>
Ipopt <http://projects.coin-or.org/Ipopt>
Interior-point optimizer for nonlinear programs.