Originally posted by: ouaigooo

Hello

I have 3 problems w/ OPL script.

PROBLEM #1 I cannot access a individual element of an float array within a tuple (see line 44 of master_simple.mod).

PROBLEM #2 When resolving (after having changed obj coef), some 'solutionValue'-calls return results that are not updated (see line 35 of master_simple.mod)

PROBLEM #3 And .printSolution() is not updated neither (see line 64 of master_simple.mod)

Below is the scripting log.

OPL version: 12.6.2 (and 12.5.1)

Thanks in advance for your help !

 [-25.51 126.35]
. set obj coef for power[D 487926000]: 25.51
. set obj coef for power[D 487929600]: -72.54000000000001
. set obj coef for power[D 487933200]: -78.79000000000001
. set obj coef for power[D 487936800]: -40.53
. set obj coef for power[D 487940400]: -14.49
 obj   : 255.1
 bound : -Infinity
. power  [10 0 0 0 0]
 . power[D 487926000] 10
 . power[D 487929600] 0
 . power[D 487933200] 0
 . power[D 487936800] 0
 . power[D 487940400] 0
// solution (optimal) with objective 255.1
// Quality There are no bound infeasibilities.
// There are no reduced-cost infeasibilities.
// Maximum Ax-b  residual             = 0
// Maximum c-B'pi residual            = 0
// Maximum |x|                        = 10
// Maximum |slack|                    = 10
// Maximum |pi|                       = 25.51
// Maximum |red-cost|                 = 104.3
// Condition number of unscaled basis = 3.0e+000
//

power = [10
         0 0 0 0];

 [-25.51 126.35]
. set obj coef for power[D 487926000]: -126.35
. set obj coef for power[D 487929600]: -73.40000000000001
. set obj coef for power[D 487933200]: -79.65000000000001
. set obj coef for power[D 487936800]: -41.4
. set obj coef for power[D 487940400]: -15.36
 obj   : 0
 bound : -Infinity
. power  [10 0 0 0 0]
 . power[D 487926000] 0
 . power[D 487929600] 0
 . power[D 487933200] 0
 . power[D 487936800] 0
 . power[D 487940400] 0
// solution (optimal) with objective 0
// Quality There are no bound infeasibilities.
// There are no reduced-cost infeasibilities.
// Maximum Ax-b  residual             = 0
// Maximum c-B'pi residual            = 0
// Maximum |x|                        = 0
// Maximum |slack|                    = 10
// Maximum |pi|                       = 0
// Maximum |red-cost|                 = 126.35
// Condition number of unscaled basis = 1.0e+000
//

power = [10
         0 0 0 0];

Hi

Pb1 : I agree there seems to be an issue. A workaround could be to avoid using arrays as a field in the tuple set and replace with a tuple set with a field scenario

Pb2&3: if after

algo.solve();

you add
    model.postProcess();

it should work better

regards

Originally posted by: ouaigooo

Thanks Alex.

Pb2&3: In my particular case, postProcess() will do some other stuff. So I would actually prefer not to call it. Is it a necessary call ? What is really bizarre is that I can still get the right values when accessing individual array elements ...

Pb1: cool suggestion, thanks. But like you say, it is a workaround. It is supposed to work the way I implemented it. I'd like to hear CPLEX's staff on this one, because I'm getting tired of workarounds all time ...

Kind regards

Hi,

about pb2&3,

let me quote documentation.

postprocess:

Used to perform postprocessing. After a solution is found, the postprocessing part of the model can be executed. The solving is controlled outside OPL, therefore this service routine has to be invoked manually whenever a new solution is available.

regards

Originally posted by: ouaigooo

Fair enough. A big thanks, Alex.