[Clp-tickets] [COIN-OR Linear Programming Solver] #25: MPS/QPS reader cannot read its own output file
COIN-OR Linear Programming Solver
coin-trac at coin-or.org
Tue Sep 16 07:21:22 EDT 2008
#25: MPS/QPS reader cannot read its own output file
------------------------+---------------------------------------------------
Reporter: nowozin | Owner: somebody
Type: defect | Status: new
Priority: major | Milestone:
Component: component1 | Version:
Keywords: |
------------------------+---------------------------------------------------
The attached prob1.qps file was created using a quadratic model in
ClpSimplex and calling writeMps.
The resulting file can be read in by BPMPD with warnings as follows:
{{{
BPMPD Version 2.21B
QP/LP optimizer
Written by Cs. Meszaros
MTA SZTAKI, Budapest
November, 1998
Sizeof BP_REAL: 8 Sizeof BP_INT: 4
Parameter file not found: bpmpd.par
Default MPS input parameters will be used
File does not exist:prob1.mps
MPS file name [*.mps] : prob1
NAME BLANK
Objective function :OBJROW
Column hashing : 2436 Row hashing : 1822
Max column search : 0 Max. row search: 0
RHS name : RHS RHS records : 1822
WARNING : wrong colname in the BOUNS:C0000002 ignored
WARNING : wrong colname in the BOUNS:C0000021 ignored
WARNING : wrong colname in the BOUNS:C0000040 ignored
WARNING : wrong colname in the BOUNS:C0000059 ignored
WARNING : wrong colname in the BOUNS:C0000078 ignored
Column name warnings ignored: 2
BOUNDS name : BOUND BOUNDS records : 126
Wrong name in QMATRIX :C0000002 C0000002
Wrong name in QMATRIX :C0000021 C0000021
Wrong name in QMATRIX :C0000040 C0000040
Wrong name in QMATRIX :C0000059 C0000059
Wrong name in QMATRIX :C0000078 C0000078
Wrong name in QMATRIX :C0000097 C0000097
Wrong name in QMATRIX :C0000116 C0000116
Number of rows : 1822
Number of columns : 2436
Free variables : 126
Number of nonzeros : 67202
Largest column length: 908 Column name :C0000057
Number of Q columns : 2436
Number of Q nonzeros : 0
Parameter file not found: bpmpd.par
Default optimization parameters will be used
MPS file is read in 0.11 sec.
Space for factors: 11042791
Process: presolv
CHEPDU: 826 columns, 0 rows removed
ROWDUP: 0 columns, 20 rows removed
Presolv done...
Process: scaling
Pass 0. Average def. 1.834768727e+00
Pass 1. Average def. 1.438823167e+00
Pass 2. Average def. 1.446040344e+00
Scaling done...
Process: aggregator
0 rows dependent
0 row/cols eliminated, 0 fill-in created.
0 nonzeros eliminated
Aggregator done...
Time for presolv, scaling and aggregator: 0.04 sec.
Number of rows : 1802
Number of columns : 1610
Free variables : 126
No. of A nonzeros : 66466
No. of Q nonzeros : 0
Largest column count : 894
Obj. scaled 1.97e-01
Number of dense rows : 0 0.0%
Largest sparse col. length : 3 0.2%
Smallest dense col. length : 39 2.2%
Number of dense columns : 126 7.8%
No. of superdense columns : 72
Ordering: Minimum local fill-in
Analyse for original supernodes : 35 col
Final supernodal columns disabled: 92 col
Hidden supernodal columns : 1 col
Building ordering list time: 0.00 sec
Symbolic factorisation time: 0.03 sec
Sub-diagonal nonzeros in aat : 10181
Sub-diagonal nonzeros in L : 71371
Total symbolic phase time: 0.03 sec
NONZEROS : 80215
OPERATIONS : 1762039
Supernodes : 3 27
Supernodal cols. : 130 185
Dense window : 107
Centrality correction Power: 2
Iterative refinement Power:12
FIRSTFACTOR TIME : 0.04 sec
It-PC P.Inf D.Inf U.Inf Actions P.Obj D.Obj
Barpar
------------------------------------------------------------------------------
0-00 2e-01 8e+01 0e+00 0 0 0 3.7834665e+03 2.7792950e+02
0e+00
1-12 3e-03 2e+00 0e+00 0 0 0 3.3090960e+03 2.8361982e+02
5e-01
2-11 7e-04 6e-01 0e+00 0 0 0 1.3541014e+03 2.8466751e+02
4e-01
3-12 7e-05 3e-01 0e+00 0 0 0 5.0868116e+02 3.2032103e+02
1e-01
4-12 2e-05 7e-02 0e+00 0 0 0 4.3272595e+02 3.5392130e+02
6e-02
5-12 7e-06 3e-02 0e+00 0 0 0 4.0383888e+02 3.6762369e+02
2e-02
6-12 3e-06 1e-02 0e+00 0 0 0 3.9242021e+02 3.7409037e+02
1e-02
7-12 8e-07 2e-03 0e+00 0 0 0 3.8550795e+02 3.7806988e+02
5e-03
8-12 2e-13 2e-07 0e+00 0 0 0 3.8189272e+02 3.8039287e+02
2e-03
9-12 3e-13 5e-08 0e+00 0 0 0 3.8118829e+02 3.8111069e+02
1e-04
10-12 4e-13 1e-08 0e+00 0 0 0 3.8115125e+02 3.8115017e+02
1e-06
11-11 3e-12 1e-09 0e+00 0 0 0 3.8115074e+02 3.8115073e+02
1e-10
Stopping Criteria:Small infeasibility and duality gap
ABSOLUTE infeas. Primal : 2.596e-12 Dual :
1.106238573e-09
PRIMAL : Relative infeas. : 1.105e-13 Objective :
3.811507366e+02
DUAL : Relative infeas. : 1.409e-11 Objective :
3.811507344e+02
Complementarity gap : 5.919e-09 Duality gap :
-5.738713979e-09
Solver time 0.24 sec.
OPTIMAL SOLUTION. OBJECTIVE = 3.811507366e+02
TOTAL SOLUTION TIME = 0.28 sec.
}}}
Clp cannot read the file in, as it usually can with QPS files. The error
is the following:
{{{
Coin LP version 1.06.00, build Sep 15 2008
command line - /opt/coin-osi-trunk20080915/bin/clp prob1.qps
At line 1 NAME BLANK
At line 2 ROWS
At line 1826 COLUMNS
At line 37177 RHS
At line 38089 BOUNDS
No match for column C0000002 at line 38092 < FR BOUND C0000002
1e+30 >
No match for column C0000021 at line 38111 < FR BOUND C0000021
1e+30 >
No match for column C0000040 at line 38130 < FR BOUND C0000040
1e+30 >
No match for column C0000059 at line 38149 < FR BOUND C0000059
1e+30 >
No match for column C0000078 at line 38168 < FR BOUND C0000078
1e+30 >
No match for column C0000097 at line 38187 < FR BOUND C0000097
1e+30 >
No match for column C0000116 at line 38206 < FR BOUND C0000116
1e+30 >
At line 38223 QUADOBJ
Problem BLANK has 1822 rows, 2436 columns and 67202 elements
There were 7 errors when importing model from ./prob1.qps
There were 7 errors on input
}}}
Usually the QPS/SIF files from http://cuter.rl.ac.uk/cuter-
www/Problems/marmes.shtml work, for example when "clp AUG3D.SIF" is
called.
--
Ticket URL: <https://projects.coin-or.org/Clp/ticket/25>
COIN-OR Linear Programming Solver <http://projects.coin-or.org/Clp>
A linear programming solver.
More information about the Clp-tickets
mailing list