Originally posted by: SystemAdmin

[a_shraideh said:]

Hellow every body,
i have strange problem,
i use  ILOG CPLEX v9, to solve MIP problem with java

this MIP sample  is generated randomly
when i solve only one sample at each execution the solution works perfectly, when i try to solve several sample or to
put my sample within a  "for loop"  to solve the different sample  i get the following results for each iteration???
(
Row '' infeasible, all entries at implied bounds.
)
a sodo code is the following
for only one sample :

1) IloCplex cplex = new IloCplex();
2) generatesample();
3) populateByRow(cplex, var, rng);
4) cplex.solve();
5) cplex.end();
this work very will


for several sample:

for(i=0;i<10;i++)<br />{
1) IloCplex cplex = new IloCplex();
2) generatesample();
3) populateByRow(cplex, var, rng);
4) cplex.solve();
5) cplex.end();
} 
I get the solution only for the first sample and then i get 9 times
Row '' infeasible, all entries at implied bounds.
Row '' infeasible, all entries at implied bounds.
Row '' infeasible, all entries at implied bounds.
Row '' infeasible, all entries at implied bounds.
Row '' infeasible, all entries at implied bounds.
Row '' infeasible, all entries at implied bounds.
Row '' infeasible, all entries at implied bounds.
Row '' infeasible, all entries at implied bounds.
Row '' infeasible, all entries at implied bounds.


thank u in advance
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: SystemAdmin

[prubin said:]

Does your populateByRow method remove all existing constraints and bounds from cplex before adding the new constraints?  If not, at each iteration after the first you are getting a mixture of old and new constraints (or maybe partially old, partially new constraints), which could be making the problem infeasible.
#CPLEXOptimizers
#DecisionOptimization

Originally posted by: SystemAdmin

[a_shraideh said:]

Hi prubin, thank you for your fast answere
yes, it is true i use populateByRow method, i tried ti use cplex.end() and cplex.clearmodel(); to remove my constraint but it didnt work , are they the right method if yes what is the problem???
thanks in advance

#CPLEXOptimizers
#DecisionOptimization

Originally posted by: SystemAdmin

[a_shraideh said:]

thanks every body,
i found the probelm, it was that every time i generate random simple i use the same variable ?? it was very stupid from my side

#CPLEXOptimizers
#DecisionOptimization