Dear Andrew,
the attached MILP model exits with:
* 5400: objval = 6.264244529e+00 infeas = 0.000000000e+00 (0)
* 5450: objval = 5.968440326e+00 infeas = 0.000000000e+00 (0)
OPTIMAL SOLUTION FOUND
Integer optimization begins...
+ 5450: mip = not found yet >= -inf (1; 0)
+ 5481: >>>>> 6.036226238e+00 >= 5.968440326e+00 1.1% (32; 0)
GLPK internal error: x <= ub; file glpios03.c, line 265
Abort
Maybe it is a scaling problem, but just wanted to let you know in case it
has triggered a bug. I have a very similar model which runs without any
problems.
With best regards
Oscar Gustafsson
data;
param inputbits := 12;
param partitionbits := 4;
param part1bits := 4;
end;
# Piecewise linear approximation
# Oscar Gustafsson, [EMAIL PROTECTED]
param inputbits, integer, > 0;
param partitionbits, integer, > 0;
param part1bits, integer;
param integerbitsoffset := 16;
param integerbitsslope := 16;
set Y := {0..(2^partitionbits-1)};
set Z := {0..(2^(inputbits-partitionbits)-1)};
param X{y in Y, z in Z} := 2^(inputbits-partitionbits)*y + z;
param fracweight := 2^(inputbits-partitionbits);
param pi := 3.14159265;
param f{y in Y, z in Z} := sin(X[y,z]/2^inputbits*pi/2);
var p{y in Y}, integer;
var q{y in Y}, integer;
var error>=0;
minimize e: error;
s.t. postive{y in Y, z in Z}: p[y]/(2^integerbitsoffset) +
z/fracweight*q[y]/(2^integerbitsslope) - f[y,z] <= error;
s.t. negative{y in Y, z in Z}: -(p[y]/(2^integerbitsoffset) +
z/fracweight*q[y]/(2^integerbitsslope) - f[y,z]) <= error;
solve;
display p;
display q;
end;
_______________________________________________
Bug-glpk mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/bug-glpk
