Originally posted by: bobby76

Hi, 

I am relatively new to tuple. I want to develop a classical job shop scheduling using OPL, not CP optimizer. I have stuck to create operation precedence constraints for operation, i.e. operation o of job j precede operation o+1.

 

Here is my script:

//This is where I stuck

 

Here is with tuple declaration:

nJobs = 6;
nMachines = 6;
nOperations = 6;
nPeriods = 17;

ops = [
 [ <5,4>, <1,3>, <4,3>, <3,2>, <0,1>, <2,2> ],
 [ <1,3>, <0,8>, <5,7>, <2,2>, <4,9>, <3,3> ],
 [ <3,1>, <4,9>, <1,9>, <0,7>, <5,5>, <2,5> ],
 [ <3,8>, <4,2>, <1,1>, <5,7>, <2,8>, <0,9> ],
 [ <1,6>, <3,2>, <4,5>, <5,5>, <0,3>, <2,1> ],
 [ <4,10>, <2,4>, <0,4>, <3,3>, <1,2>, <5,3> ]
];

 

Has anybody experienced in building job shop model, please help me. Your help is appreciated.

 

Best regards,

Bobby

 

 

 

 

 

 

 

#DecisionOptimization
#OPLusingCPLEXOptimizer

Hi,

you have an example at https://www.ibm.com/developerworks/community/forums/html/topic?id=8ebacb94-8758-444b-a5c2-b9b137b69370&ps=25

But CPO is quite good for jobshop. So I wonder why you do not use that.

regards

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: bobby76

Hi Mr Fleischer,

Thanks for your information. I will run your code with instances from Fisher and Thompson 6x6 instance (ft06) with known optimal Cmax = 55.

Best regards,

 

Bobby

 

 

 

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: Ladas

Hi,

even though the topic is solved, I want to ask very similar question, so I put it here.

I am also dealing with JSSP and I wanna use CPLEX (for comparison with CP). I also looked at the examples in ibm ilog cplex OS and modified it for my case (file attached). I use special case, when some machine can be utilized multiple times during one job and every job has different number of operations. But I guess it is poorly specificated because it doesn't give me any result even after almost 2 hours! When I put 0 instead of objective function, after 6 seconds I got poor result (there are downtimes of the machines). But I don't know any other way, despite the fact that there should be a way (since JSSP is among CP also MIP problem)...

If you have any idea how to make it solvable, I will really appreciate that!

Reagards!

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: bobby76

Hi Mr. Ladas,

 

I have test the Mr. omidi model using ft06 benchmark and the result is not good. 

 

// solution (feasible relaxed sum of infeasibilities) with objective 158

whereas the optimal makespan indeed 55.

 

I think you the problem lies on c04 in Mr. omidi model because it generates conflict.

 

Best regards,

 

Bobby

 

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: Ladas

Thanks for answer.

But I meant, I don't even need that x[][][] and A[][][] because I am not dealing with the problem of multiple machines for a single operation. Furthermore I do not have the same number of operations for every job.

I tried to fit my case to that of Mr. omidi, But it seems impossible (since number of operations within one job is greater than number of machines)...

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: bobby76

Hi Mr. Ladas,

 

I think your problem is mix of a flexible job shop and flow shop. I still think you need x[jobs][operations][mhacines] and A[jobs][operations][machines] or tuple to do this. For example, you need to distinct between operation and machine. So, for example 2 jobs with different number of operations. The operations are performed by 2 machines (i.e. a machine can process more than one operation).

Job 1 has 3 operation with precedence rule (i.e. operation 2 can not be performed until operation 1 is finished)

Job 2 has 2 operation with precedence rule (i.e. operation 2 can not be performed until operation 1 is finished)

 

Processing time data                           Routing data 

             O1  O2   O3                              O1       O2       O3

Job 1:   4     3      2                    Job 1:  2           1          2

Job 2:  2      4      0                    Job 2:  1           2

 

Let p[i][j] denotes the processing time of job i of operation j. 

Let A[i][j][k] be a binary parameter, i.e. 1 if job i operation j is processed on machine k, 0 otherwise.

Your data in OPL will be like this:

p =

[  [4 3 2]  //Job 1  

  [2 4  0]  //Job 2

];

A = 

[

[  [ 0 1] [1 0] [0 1]  ] //for job 1

[  [1 0] [0 1] [0 0]   ] //for job 2  

];

 

Maybe use tuple for data structure. The rest constraint must be modified.

 

best regards,

 

 

 

 

 

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: Ladas

Thanks a lot for the answer. Maybe I spent some time to absorb it :) but even though I have to ask about things I still do not understand in Mr Omidi example. e.g. 

ord(jobs,k) != ord(jobs,j) && ord(jobs,j)>0

it means that job k is always before job j but not equal? Or why can I use j != i and j > 0?

And what was actually

A[jobs][machines][machines]

and

x[jobs][jobs][machines]

for?

#DecisionOptimization
#OPLusingCPLEXOptimizer