30 Jan 2012 ==============================
Mixed Integer (QC)QP Benchmark
==============================
H. Mittelmann (mittelmann@asu.edu)
The MPS-datafiles are in plato.asu.edu/ftp/miqp/ and the AMPL files in plato.asu.edu/ftp/ampl_files/miqp_ampl/ and egon.cheme.cmu.edu/ibm/files/
The following codes were run in default mode on a 2.66GHz Intel Core2 Quad, For accuracy reached, see logfiles. In the last columns are results for the QCQPs obtained by rewriting the QPs as: min t, subject to quadratic obj <= t plus constraints. SCIP does this transformation itself. SCIP uses relgap=0;
CPLEX-12.4.0.0: CPLEX (QCQP with 12.3)
Bonmin-1.5.1: projects.coin-or.org/Bonmin (Bonmin: hybrid algorithm. with Cbc)
Couenne-0.4.1: projects.coin-or.org/Couenne
FilMINT: currently only at NEOS, (run locally)
GUROBI-4.6.1: gurobi.com
SCIP-2.1.0: scip.zib.de (with CPLEX and IPOPT)
XPRESS-7.2.1: XPRESS
Times given are user times in seconds. A time limit of 10,800 seconds was imposed. "t" time limit exceeded, "f" fail, "c" problem convex.
QP
==
Scaled geometric means of runtimes
60 192 2.70 261 1 14.5 2.30
========================================================
problem Bonm Couen CPLEX FilMI GUROBI SCIP XPRESS
========================================================
iair04 1521 t 552 t 34 338 24
iair05 689 t 372 f 42 250 29
ibc1 t 10200 263 t 19 28 8
ibell3a c 39 302 3 7229 1 19 10
ibienst1 c 1323 945 263 t 8 27 17
icap6000 c 1165 t 8 2567 1 16 1
icvxqp1 c t f 1 f 4 t 38
ieilD76 370 11 29 1270 78 61 27
ilaser0 c f f 18 f 55 f t
imas284 38 t 7 221 3 16 6
imisc07 177 t 39 7865 12 54 27
imod011 c 9456 f 62 f 1 481 1
inug06-3rd c t t 565 t 141 608 211
inug08 c 65 267 5450 f 1 21 1
iportfolio c t f 33 t 84 t t
iqap10 398 f 188 4438 204 501 102
iqiu c 506 4400 37 2014 26 83 19
iran13x13 269 f 17 2592 14 32 7
iran8x32 138 f 4 5561 3 15 2
isqp c f t 1335 f 1021 t t
isqp0 c f t 26 f 232 t 287
isqp1 c f t 50 f 133 t 1628
iswath2 1185 f 121 t 9 338 21
itointqor c t f 13 f 64 f 102
ivalues t 550 f t f f f
========================================================
QCQP
====
Scaled geometric means of runtimes
8.82 26.86 2.46 30.0 1
==================================================
problem Bonmin Couenne CPLEX FilMINT XPRESS
==================================================
iair04 3692 t 38 f 24
iair05 2436 t 25 f 33
ibc1 t t 2668 160 9
ibell3a c 39 284 2 932 7
ibienst1 c 1755 1753 140 t 16
icap6000 c 1175 t 14 1599 1
icvxqp1 c t f f t 36
ieilD76 302 4804 9 1478 f
ilaser0 c f f t f t
imas284 38 2855 85 696 5
imisc07 214 t 58 t 26
imod011 c t f 5557 f 2
inug06-3rd c t t 1982 t 163
inug08 c 317 713 766 t 1
iportfolio c t t t f t
iqap10 648 7655 324 t f
iqiu c 315 5266 129 5647 19
iran13x13 204 f 66 2196 7
iran8x32 95 55 27 6709 3
isqp c f t t f t
isqp0 c f t 6217 f 290
isqp1 c f t t f 1634
iswath2 1339 t f t f
itointqor c f f 5392 f 116
ivalues t f f 5895 f
CLay0203M c 2 5 1 11 1
CLay0204M c 8 8 1 40 3
CLay0205M c 32 56 2 64 42
CLay0303M c 3 10 1 14 1
CLay0304M c 8 47 2 77 23
CLay0305M c 36 73 3 78 84
=================================================
Problem Statistics
===========================================================
problem rows col integer nonzero quad elts$
===========================================================
iair04 823 8904 8904 81869 8903
iair05 426 7195 7195 59316 7196
ibc1 1913 1751 252 276843 241
ibell3a 104 122 60 390 59
ibienst1 576 505 28 2185 27
icap6000 2171 6000 6000 24238 22507
icvxqp1 5000 10000 10000 14998 29984
ieilD76 75 1898 1898 21009 1897
ilaser0 1000 1001 151 3000 3231
imas284 68 151 150 9782 149
imisc07 212 260 259 8620 258
imod011 4480 10957 97 29838 29984
inug06-3rd 3972 2886 2886 20232 26023
inug08 912 1632 1632 8304 2984
iportfolio 202 1200 967 202200 200
iqap10 1820 4150 4150 20810 4149
iqiu 1192 840 48 3744 47
iran13x13 195 338 169 1014 168
iran8x32 296 512 256 1536 255
isqp 250 1000 50 3552 17801
isqp0 250 1000 50 3552 17801
isqp1 250 1000 100 3552 17801
iswath2 483 6404 2213 34085 2212
itointqor 1 50 50 48 115
ivalues 2 202 202 404 3620
-----------------------------------------------------------
CLay0203M 54 30 18 162 6
CLay0204M 90 52 32 272 8
CLay0205M 135 80 50 410 10
CLay0303M 66 33 21 201 6
CLay0304M 106 56 36 324 8
CLay0305M 155 85 55 475 10
===========================================================
$ in one half