Using openSUSE Linux 13.2 and Minizinc 0.9.9/2.0.6, in the (bad) code:
array[int] of set of 1..10: thesets = [{1,2},{3,5},{4,6}];
set of 1..10: sett = {5,6,9};
var 1..10: y;
var set of 1..10: z;
constraint y == min(thesets[3]);
constraint z == lb(thesets[3]);
solve satisfy;
minizinc responds with a message indicating that the lb() builtin is missing.
I think this is incorrect, the lb() function is there, but the code referring to it is bad.
A colleague in Prof. Stuckey's online course points out that if the lb() is used correctly to refer to a non-parameterized set then it works, and I confirm this is correct.
A second colleague points out that there are issues with the ub() function as well.
Please see original comments in thread:
https://class.coursera.org/modelingoptimization-001/forum/thread?thread_id=306
Using openSUSE Linux 13.2 and Minizinc 0.9.9/2.0.6, in the (bad) code:
array[int] of set of 1..10: thesets = [{1,2},{3,5},{4,6}];
set of 1..10: sett = {5,6,9};
var 1..10: y;
var set of 1..10: z;
constraint y == min(thesets[3]);
constraint z == lb(thesets[3]);
solve satisfy;
minizinc responds with a message indicating that the lb() builtin is missing.
I think this is incorrect, the lb() function is there, but the code referring to it is bad.
A colleague in Prof. Stuckey's online course points out that if the lb() is used correctly to refer to a non-parameterized set then it works, and I confirm this is correct.
A second colleague points out that there are issues with the ub() function as well.
Please see original comments in thread:
https://class.coursera.org/modelingoptimization-001/forum/thread?thread_id=306