Originally posted by: 88Simon88

Hello everybody,

Merry Christmas.

I am running a model, but it takes a lot of time. So, I use an algorithm to find a feasible solution to the problem. Now I am going to feed it to the OPL and run the model and see how much the feasible solution can reduce the runtime. Could you please explain me how can I do that? I would appreciate you if you could provide a sample code.

Best regards

Hi

have you had a look at

https://www.ibm.com/developerworks/community/forums/html/topic?id=75766584-44c5-4400-977f-03363feea776&ps=25

?

regards

Originally posted by: 88Simon88

Thanks Alex for your reply.

I read the example which you have provided and I somehow got the method, but I am not still sure. Please put me right if I am wrong: 

In order to provide a solution for CPLEX, first I have to provide the decision variables to the model and then solve it. right?

What are the arguments of the below code? decision variables? 

cplex.addMIPStart(thisOplModel.nbBus,thisOplModel.naiveSolution);

If yes, Can I only provide the decision variable (for example just to mention the number of buses with 30 and 40 seats) to the model instead of solving it (instead of below code) and then feed it to the CPLEX?

regards

Hi,

you do not need to solve before. You just need to provide values.

See https://www.ibm.com/support/knowledgecenter/ru/SSSA5P_12.8.0/ilog.odms.ide.help/refjsopl/html/IloCplex.html#addMIPStart

for the parameters for addMiIPStart

You may also find a warmstart example in CPLEX_Studio128\opl\examples\opl\warmstart

And also have a look at https://www.ibm.com/developerworks/community/forums/html/topic?id=c3f299b6-46d5-4d52-8371-d8cd81ecbd4e&ps=25

regards

Originally posted by: 88Simon88

Hi,

I read the example for a warm start and a question came up to my mind, If I have 4 decision variables, Is it enough to provide only one decision variable for the warm start? For example, in the OPL example for the warm start, if you write "y" instead of "x" in the following line, the result is the same.

vectors.attach(opl2.y,opl2.values);

best regards

Hi,

in the example warmstart, if you write

 vectors.attach(opl2.y,opl2.values);

instead of

 vectors.attach(opl2.x,opl2.values);

Then nothing much happens since y values are set by

forall( j in r )
    ctEqual:
      y[j] == j;

regards