Originally posted by: harshniu

Here is my code for the scheduling of n jobs on m machines. I am describing the problem just below the code.

//data
{int}total=...;
int m=...;
int n=...;

range job=0..n;
range mac=1..m;
//int h=m*n*(n+1);

int t;

tuple timeType{
int precedingjob;
int succeedingjob ;
int machine;
int time;}

timeType timestotal=...;

//variables
dvar int+ cjob;
dvar int+ cmax;
dvar boolean xjobjobmac;
//objective
minimize cmax;
//constraints
subject to {
forall(j in job)
onejobprecedesanother:
sum(i in 0..n, k in mac :i!=j)x[i][j][k] == 1;

forall(h in job, k in mac)
balanceconstraint:
sum(i in 0..n:i!=h)x[i][h][k] - sum(j in 0..n:j!=h )x[h][j][k] == 0;

forall(i in 0..n,j in 1..n,k in 1..m :t ==((((n+1)*j)+i-n)*(k-1)))
//p=((7*j)+i-6)*(k - 1);

jobcompletiontime:
+ **c[j]>=c[i]+sum(k in 1..m) x[i][j][k]*(times[t].time) + 10000*(sum(k in 1..m)x[i][j][k]-1);+

forall(k in mac)
firstjob :
sum(j in 0..n) x[0][j][k]==1;

timeatstart:
c[0]==0;
forall(j in job)
cmaxconstraint:
cmax>= c[j];}

As you can see I am using a tuple and I have linked my data file to and excel file to input the data. I have italicised the constraint which is giving the problem. The element(times[t].time)takes the first value from the excel sheet and does not follow the expression t ==((((n+1)*j)+i-n)*(k-1))). Thereby, multiplying the same vale say 116 * 5 for 5 job problem.
Please help me out or suggest a way to handle 3 dimensional arrays.

Thanks
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: far123

According to my experience with ODME3.3, it may not be possible to use excel file to handle 3 dimensional arrays. You can use .dat file instead.
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: harshniu

The linking of tuple to excel file enables me to access the multi-dimensional data. There is no problem with that. The problem is 'how to read multiple lines from the excel spreadsheet. e.g. D1 :D36. I am using that counter 't' to do that. But it does not take values dynamically.

And till now I haven't seen a .dat file containing data for 3-dimensional array. Can you explain how to do that?
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: far123

The below 2*4*3 array is an example of 3 dimensional array in .dat file.

[[[1,0,1] [1,1,0] [1,0,1] [0,0,1]]
 [[1,0,1] [1,1,0] [1,0,1] [0,0,1]]]

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: harshniu

Thanks , that does answer my question about 3-D in .dat file. But the reason i shifted to tuples is because I have like hundreds of data files in .txt format & it is impossible to convert each one of them into the desired .dat format. So, I converted them into a column in a spreadsheet writing a .cpp code. Now I have the files in the excel format. So, can you help me with my original problem?
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

There is a FAQ to read a 3-D array from Excel: http://www-01.ibm.com/support/docview.wss?rs=0&context=SSCMS55&uid=swg21401340&loc=en_US&cs=utf-8&cc=us&lang=all but this FAQ assumes you already know the size of the array.
If the issue you're facing is more related to the range that is only known at runtime, maybe you could use named range in your Excel spreadsheet ?
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: UDOPS

Correct me if I am wrong, but the constraint definition appears wrong and this issue may have nothing to do with data files. In the "for" statement, you define ranges for i, j, k; but only list t as a limiting condition. If you want to read from the tuple set, shouldn't it be "for (<i,j,k,t> in timestotal: t ==((((n+1)*j)+i-n)*(k-1))) " ?

By the way, I think this would be a lot more efficient if you separated the set in your excel file first, rather than parse it in OPL, since the test is a static condition.
#DecisionOptimization
#OPLusingCPLEXOptimizer