bug-glpk
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Bug-glpk] Bug?


From: Johan Löfberg
Subject: [Bug-glpk] Bug?
Date: Thu, 27 Nov 2003 12:50:49 +0100

Experiencing a lot of strange behavior when turning on/off presolve/scaling
in various ways

(The problems are genrated during a b'n'b loop, so they are probalby a bit
strange with a lot of redundancy etc)

Win XP, glpk4.2, gcc 3.3.1

Works

./glpsol --lpt --nopresol --noscale outpb.lp

lpt_read_prob: reading LP data from `outpb.lp'...
lpt_read_prob: 15 variables, 115 constraints
lpt_read_prob: 190 lines were read
lpx_adv_basis: size of triangular part = 115
      0:   objval =   7.156220499e+01   infeas =   1.000000000e+00 (0)
     39:   objval =  -2.746063571e+01   infeas =   2.823863578e-18 (0)
*    39:   objval =  -2.746063571e+01   infeas =   9.696687897e-13 (0)
*    48:   objval =  -1.000000000e+02   infeas =   2.784061870e-11 (0)
OPTIMAL SOLUTION FOUND
Time used:   0.0 secs
Memory used: 0.1M (85208 bytes)


Changing scaling

./glpsol --lpt --nopresol --noscale outpb.lp
lpt_read_prob: reading LP data from `outpb.lp'...
lpt_read_prob: 15 variables, 115 constraints
lpt_read_prob: 190 lines were read
lpx_adv_basis: size of triangular part = 115
      0:   objval =   7.156220499e+01   infeas =   1.000000000e+00 (0)
     41:   objval =   5.000000000e+01   infeas =   1.098117978e-19 (0)
*    41:   objval =   5.000000000e+01   infeas =   9.250462064e-15 (0)
*    45:   objval =  -4.900761058e+01   infeas =   2.154115959e+00 (0)
lpx_simplex: numerical instability (primal simplex, phase II)
     45:   objval =  -4.900761058e+01   infeas =   1.000000000e+00 (0)
     46:   objval =  -4.972478675e+01   infeas =   4.966579444e-09 (0)
*    46:   objval =  -4.972478675e+01   infeas =   1.069858804e-08 (0)
Assertion failed: ap[q] != 0.0; file glpspx2.c; line 1253

/johan löfberg

Attachment: outpb.lp
Description: Binary data


reply via email to

[Prev in Thread] Current Thread [Next in Thread]