Originally posted by: user1234567

Hello, i have some erros in the results. I have an error with the s and c decision variables. It shows very big values.

In each row, i should get one value of c and s greater than 0. The 2 other values should be equal to 0 because each operation can be processed by only one machine.

But, instead of getting 0, i get big values of c and s.

Please help me to correct these errors. Thank you

Hi,

some of your s do not have bounds so when you do

minimize staticLex(max(i in Jobs, o in Ops: o.pos==jlast[i]) endOf(ops[o]),-sum(i in Jobs,j in Oper, k in Mchs) s[i][j][k]);

you will get high values.

dvar int s[i in Jobs][j in Oper][k in Mchs] in 0..1000;

will limit those values and let you see which s do not have bounds : the ones that are 1000 in the solution.

Regards

Originally posted by: user1234567

Okay if i write

dvar int s[i in Jobs][j in Oper][k in Mchs] in 0..1000;

 

,

i will get the same problem:

s = [[[0 0 0]
                 [0 2 777]
                 [449 8 0]
                 [11 21 989]]
             [[438 545 722]
                 [247 31 640]
                 [264 177 7]
                 [49 400 64]]
             [[312 193 786]
                 [367 815 707]
                 [709 950 939]
                 [704 231 918]]];

This is impossible because o21 starts here on machine 2 and 3 at the same time (according to s values); on machine 2 it starts at 2, and on machine 3 it starts at 777. Whereas, o21 should be processed on machine 2; starts at 2 and ends at 8 (same as in journal de script). Same errors in the c decision variable!  And if i use these decison variables in the contraints later i will get an incorrect or infeasible result. 

Originally posted by: user1234567

For example if i add this constraint :

 forall (i1,i2 in Jobs, j1,j2 in Oper, o1,o2 in Ops: o2.pos==o1.pos-1 && o1.id==j1 && o2.id==j2, k in Mchs)
 c[i1][j1][k] >=  sum (k in Mchs) c[i2][j2][k] + sum (i in Jobs, j in Oper, k in Mchs) y[i1][j1][i2][j2][k] * p[i1][j1][k];

i get no result !  because of c decision variable which contains false results.

Well your question was about why you get big values. You get big values because some of your decision variables are not bound.

In the problem browser you may see ctc1 and ctc2

regards