VIsemiLSQRnew2

Description:

VIsemiLSQRnew2-code is a Fortran 90 software designed to solve large scale variational inequality problems using the generalisation of the Inexact Newton method applied to a semismooth nonlinear system (see [3]). This approach requires a sparse and large linear system to be solved at each step. I n particular we have used LSQR method  [2], combined by a convenient preconditioner (a variant of Incomplete LU-factorization [1]) . Furthemore we have introduced an adaptive stopping rule (see [3]).

Authors:

Instructions and documentation

We have interfaced the Fortran code by AMPL.

 

1)NLP TEST PROBLEMS

The file seminlp contains the execution file obtained using double precision on DEC ALPHA 21264 EV6/7633Mhz  with 512MB of RAM, running Linux RH7.2 ALPHA operation system.

  1. Download the seminlp
  2. Line-command:              ./seminlp  file.nl  [-AMPL] 

 

An Example:

1)Download the example  (torsion4.mod, torsion4.nl ) 

2)Line command:              ./seminlp  torsion4.nl  [-AMPL] 

3)Download outputfile:  out_torsion4                      

 

2)MCP TEST PROBLEMS

The file semimcp contains the execution file obtained using double precision on DEC ALPHA 21264 EV6/7633Mhz  with 512MB of RAM, running Linux RH7.2 ALPHA operation system.

  1. Download the semimcp
  2. Line-command:              ./semimcp  file.nl  [-AMPL] 

 

An Example:

1)Download the example (bersek.mod, bersek.nl). 

2)Line command:          ./semimcp  bertsek.nl  [-AMPL] 

3)Download outputfile: out_bertsek

 

References
  1. B.A. Murtagh and M.A. Sauder, MINOS 5.0 user's guide.Stanford Uniersity, Stanford, California,(1983).
  2. C.C.Paige and M.A.Saunders, LSQR: An Algorithm for Sparse Linear Equations and Sparse Least Squares, ACM Transactions on Mathematical Software 8(1), (1982), p.43-71 
  3. V. Ruggiero, F. Tinti, A preconditioner for solving large scale variational inequality problems by a semismooth inexact approach, rapporto tecnico n.69 del Dipartimento di Matematica Università di Modena, 2005;